• Save
Paper 1 Phd Course Work- Research Methodology Exam
Upcoming SlideShare
Loading in...5
×
 

Paper 1 Phd Course Work- Research Methodology Exam

on

  • 1,104 views

Paper 1 Phd Course Work- Research Methodology Exam

Paper 1 Phd Course Work- Research Methodology Exam

Statistics

Views

Total Views
1,104
Views on SlideShare
1,097
Embed Views
7

Actions

Likes
0
Downloads
0
Comments
0

1 Embed 7

http://www.slideee.com 7

Accessibility

Categories

Upload Details

Uploaded via as Adobe PDF

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

Paper 1 Phd Course Work- Research Methodology Exam Paper 1 Phd Course Work- Research Methodology Exam Document Transcript

  • Paper1: Research Methodology Exam Sheet   Shivananda  R  Koteshwar,  PhD  Research  Scholar,  Bangalore  University   1   Paper  1  Exam  Sheet   Research  Methodology  and  Statistics   Shivananda  R  Koteshwar   TITLE:  A  Study  on  Pragmatic  Approaches  and   Quality  Initiatives  for  Enhancing  Teachers’   Caliber  in     Post  Graduate  Institutes  offering  MBA   Programme  under  Bangalore  University       Under  the  Guidance  of       Dr.  T.V.  Raju   Director,  RV  Institute  of  Management,  Bangalore   CANARA  BANK  SCHOOL  OF  MANAGEMENT  STUDIES   BANGALORE  UNIVERSITY  
  • Paper1: Research Methodology Exam Sheet   Shivananda  R  Koteshwar,  PhD  Research  Scholar,  Bangalore  University   2   RESEARCH METHODOLOGY Life cycle of Research.................................. 3 Scientific Research..................................... 3 Research Process........................................ 4 Research Report......................................... 5 Good measurement characteristics........................ 7 Research Problem........................................ 8 Hypothesis.............................................. 8 Case Study............................................. 11 Sampling............................................... 11 Data Preparation Process............................... 12 STATISTICS   Characteristics of a statistical data.................. 13 Arithmetic Mean........................................ 13 Median................................................. 14 Mode................................................... 14 Standard Deviation and Variance........................ 14 Coefficient of Variation............................... 15 Range and Coefficient of Range......................... 15 Trend Analysis (Straight Line Analysis)................ 15 Standard Normal Curve (SNC)............................ 16 Non parametric test – (χ2) kai2 test ................... 16 ANNOVA – Analysis of Variance.......................... 17 Coefficient of Correlation............................. 20 Regression............................................. 20 Small Sample Test...................................... 21 IMPORTANT QUESTIONS  
  • Paper1: Research Methodology Exam Sheet   Shivananda  R  Koteshwar,  PhD  Research  Scholar,  Bangalore  University   3   Life cycle of Research • Hypothesis,   Prediction,   Formulation   of   question,   Sampling,   Experimentation,   Observation,   Recording,   Measurement,   Analyzing,   Formulation,  Testing,  Modification  and  Conclusion     Types of Research (PAD DEEA) (ASHE) • Either  based  on  Intent  or  based  on  method   • Intent  Based:  Pure,  Applied,  Exploratory,  Action,  Descriptive,  Diagnostic,   Evaluation     • Method   Based:   Experimental,   Analytical/Statistical,   Historical,   Survey/Fact  Finding       o Pure:  Undertaken  for  the  sake  of  knowledge  without  any  intention   to  apply  it  in  practice.  Aims  at  extension  of  knowledge   o Applied:  Problem  oriented  and  action  directed.  Gives  conceptual   clarity     o Exploratory:   Formulative   Research.   Study   of   an   unfamiliar   problem   about   which   the   researcher   has   little   or   no   knowledge.   Usually  takes  the  form  of  a  pilot  study   o Descriptive:   Fact   finding   investigation.   More   specific   than   exploratory  research.     o Diagnostic:   Similar   to   descriptive   but   with   a   different   focus.   Directed   towards   discovering   what   is   happening,   why   is   it   happening  and  what  can  be  done  about   o Evaluation:   Type   of   Applied   research.   Made   for   assessing   the   effectiveness  of  social  or  economic  programmes  implemented     o Action:  It’s  a  type  of  evaluation  study.  It  is  a  concurrent  evaluation   of   an   action   programme   launched   for   solving   a   problem   for   improving  an  existing  situation   o Experimental:  Assessing  the  effects  of  a  particular  variables  on  a   phenomenon  by  keeping  the  other  variables  constant  or  controlled   o Analytical:   Known   as   Statistical   Method.   System   of   procedures   and  techniques  of  analysis  applied  to  a  quantitative  data   o Historical:  Study  of  past  records.  Tries  to  discover  the  trends  in   the  past   o Survey:   Fact   finding   study.   Purpose   is   to   provide   information,   explain   phenomenon   to   make   comparisons   and   concerned   with   cause  and  effect  relationships     Scientific Research • A   method   or   procedure   consisting   of   systematic   observation,   measurement,   and   experiment,   and   the   formulation,   testing,   and   modification  of  hypotheses”   • Requires  replication,  external  review  and  data  recording  &  sharing   • The  key  elements  of  scientific  research  (articles  of  faith)  are      
  • Paper1: Research Methodology Exam Sheet   Shivananda  R  Koteshwar,  PhD  Research  Scholar,  Bangalore  University   4   o Ethical  neutrality  (Eliminate  personal  opinion)   o Reliance  on  empirical  Evidence     o Use  of  relevant  concept   o Commitment  of  Objectivity   o Generalization   o Validity  &  Reliability   o Logical  Reasoning  process   • Scientific  research  method  is  inquiry  based  on  empirical  and  measurable   evidence  subject  to  specific  principles  of  logic  reasoning   • Effective  Methodology:  Question    Observe    Hypothesis    Prediction     Test    Analyze    Interpret    Publish    Retest   Research Process • Research  Area/Theme/Problem/Idea   • Tentative  hypothesis   • Literature  Review   • Research  Title/Topic   • Research  Questions   • Research  Proposal   o Need  for  study   o Limitation  of  Research   o Scope  of  Research   o Budget   o Responsibilities  and  Obligations  of  stake  holders   o Place  and  Period  of  study   • Research  Proposal  Approval   • Objectives   • Hypothesis   • Operational  definition   • Research   Method/   Research   Design   (Type,   Purpose,   Timeframe,   Scope   and  environment)   o Research  Type    Experimental,  Historical  and  Inferential  Designs    Exploratory,  Descriptive  and  Causal  Designs    Experimental  and  Post  facto    Historical  method,  Case  study,  Clinical  Study    Sample  Surveys,  Field  studies,  Experiments  in  field  settings,   Laboratory  experiments    Exploratory,  Descriptive,  Experimental  studies    Exploratory,  Descriptive,  Casual    Experimental,  Quasi-­‐Experimental  Designs    True   Experimental,   Quasi-­‐Experimental   and   Non   experimental  designs    Experimental,   Pre-­‐Experimental,   Quasi-­‐Experimental   designs,  Survey  Research   o Research  question  or  purpose   o Research  timeframe   o Data  Collection  Design  
  • Paper1: Research Methodology Exam Sheet   Shivananda  R  Koteshwar,  PhD  Research  Scholar,  Bangalore  University   5    Variables    Data  collection  methods   o Sampling  Design    Sample  Population  and  Sampling  Size    Sample  Distribution  Decision    Sampling  Method/Technique    Sampling  Unit/Frame   o Instrument  Development    Introduction  and  Instructions  for  participants    Target  Questions  (AIM)   • Administrative  Questions   • Investigative  questions   • Measurement  Questions    Preliminary  Analysis  plan   o Pilot  testing   • Data  collection  and  preparation   • Data  Analysis     o Findings  (Testing  of  hypothesis)   o Interpretation  and  Conclusions   • Report  writing  /  Research  Reporting     Note:   Research   type   is   categorized   based   on   the   different   perspectives   from   which  any  given  study  can  be  viewed.  They  are:   • The  degree  of  formulation  of  the  problem  (Exploratory  or  Formalized)   • The   topical   scope-­‐breadth   and   depth   of   the   study   (Case   or   statistical   study)   • The   research   environment   (Field   Setting/Survey   or   laboratory   experiment)   • The  time  dimension  (one-­‐time  or  longitudinal)   • The  mode  of  data  collection  (Observational  or  survey)   • The  nature  of  relationship  among  variables  (Descriptive  or  casual)   Research Report   Broad  Divisions   Individual  Sections   Title  of  Report   Table  of  Contents     Preliminary  material   Abstract/Synopsis   Introduction   Literature  Review   Methodology   Results   Discussion   Conclusion   Body  of  report   Recommendations   References  or  Bibliography  Supplementary   material   Appendices  
  • Paper1: Research Methodology Exam Sheet   Shivananda  R  Koteshwar,  PhD  Research  Scholar,  Bangalore  University   6   Levels of Measurement / Measurement Scales (NOIR)  (ODO)   • Nominal:   Consists   of   assigning   numerals   or   symbols   to   different   categories  of  a  variable.  They  are  just  like  labels  and  have  no  quantitative   value.  E.g.:  Male  and  Female  applicants  of  a  MBA  program   • Ordinal:  Persons  or  objects  are  assigned  numerals,  which  indicate  ranks   with  respect  to  one  or  more  properties  either  in  ascending  or  descending   order.   E.g.:   Ranking   of   individual   based   on   socio-­‐economic   class,   which   might  be  a  combination  of  income,  education,  occupation  and  wealth   • Interval:   It’s   ranking   with   equality   in   distance.   So   it’s   not   possible   to   multiply  or  divide  the  numbers  on  an  interval  scale.  E.g.:  The  centigrade   temperature   gauge.   A   temperature   of   50degrees   is   exactly   10   degrees   hotter  than  40  degrees  and  10  degrees  cooler  than  60  degrees   • Ratio:   This   has   absolute   zero   point.   Since   there   is   natural   zero,   it   is   possible  to  multiply  and  divide  the  numbers  on  a  ratio  scale.  E.g.:  Height,   weight,  distance  and  area   MEASUREMENT   ORDER   DISTANCE   ORIGIN   STATISTICAL   TOOL  USED   SCALES  USED   Nominal   NO   NO   NO   None   Simple  Category,   Multiple  choice,   Single  Response,   Multiple  Choice,   Multiple   response,  Graphic   Rating  scale   Ordinal   YES   NO   NO   Median,  Rank   order   correlation   coefficient   Stapel  Scale   Interval   YES   YES   NO   Standard   Deviation,   Product   Moment   correlation,  “t”   tests,  “F”  tests   Likert  scale   summated  Rating,   Semantic   Differential  Scale,   Numerical  Scale,   Multiple  rating   list  scale,  Staple   scale,  Graphic   Rating  scale   Ratio   YES   YES   YES   Standard   Deviation,   Product   Moment   correlation,  “t”   tests,  “F”  tests,   Geometric   Mean,   Coefficient  of   variation   Constant  sum   scale,  Graphic   Rating  Scale  
  • Paper1: Research Methodology Exam Sheet   Shivananda  R  Koteshwar,  PhD  Research  Scholar,  Bangalore  University   7   • The  measurement  scales,  commonly  used  in  marketing  research,  can  be   divided  based  on  number  of  dimensions:     o Comparative  and  Non  comparative  scales    Comparative   scales   involve   the   respondent   in   signaling   where   there   is   a   difference   between two   or   more   producers,   services,   brands   or   other   stimuli.   Examples   of   such   scales   include;   paired   comparison,   dollar   metric,   unity-­‐sum-­‐gain  and  line  marking  scales.      Non-­‐comparative   scales,   described   in   the   textbook,   are;   continuous   rating   scales,   line-­‐marking   scales,   itemized   rating  scales,  semantic  differential  scales  and  Likert  scales.   o Uni-­‐dimensional  Scale  and  Multi-­‐dimensional  scale   o Balanced  or  unbalanced  scale   o Forced  or  Un  forced  choice  scale   o Simple  Category  scale  (Dichotomous  scale),  Multiple  choice  single   response   scale   and   Multiple   choice   Multiple   response   scale   (multiple  choice  scale)   o Likert   scale   (Summated   rating   scale)   and   Semantic   Differential   Scale  (SD  Scale)   SCALE   MEASUREMENT   Simple  Category  Scale   Nominal   Multiple  Choice  Single  Response  Scale   Nominal   Multiple  Choice  Multi  Response  Scale   Nominal   Likert  Scale  summated  rating   Interval   Semantic  Differential  Scale   Interval   Numerical  Scale   Ordinal  or  Interval   Multiple  Rating  List  scale   Interval   Constant  Sum  Scale   Ratio   Stapel  Scale   Ordinal  or  Interval   Graphic  Rating  Scale   Ordinal  or  Interval  or  Ratio       Good measurement characteristics • Uni-­‐dimensionality   • Linearity   • Validity:  (ConPreCon)     o Validity   refers   to   how   effective   an   instrument   is   in   measuring   a   property  it  intends  to  measure.     o Three   types   of   validity   are   Content   Validity   (Face   Validity   and   Sampling  Validity),  Predictive  Validity  and  Construct  Validity   o Content   Validity-­Face   Validity:   Subjective   evaluation   of   a   measuring  scale.  E.g.  a  researcher  may  develop  a  scale  to  measure   consumer  attitude  towards  a  brand  and  pre-­‐test  the  scale  among  a   few   experts.   If   the   researchers   are   satisfied,   the   researcher   may   conclude  that  the  scale  has  face  validity   o Content   Validity   –   Sampling   Validity:   Refers   to   how   representative  the  content  of  the  measuring  instrument  is.  E.g.  If   attitude  is  the  characteristic  being  measured,  its  content  universe  
  • Paper1: Research Methodology Exam Sheet   Shivananda  R  Koteshwar,  PhD  Research  Scholar,  Bangalore  University   8   may  comprise  statements  and  questions  indicating  which  aspects   of  attitude  need  to  be  measured.  This  is  also  based  on  judgment   o Predictive   Validity:   Refers   to   the   extent   to   which   one   behavior   can  be  predicted  based  on  another.  E.g.  In  the  case  of  admission   test  designed  for  prospective  MBA  students,  the  predictive  validity   of   the   test   would   be   determined   by   the   association   between   the   scores  on  the  test  and  the  grade  point  average  secured  by  students   during  the  first  semester  of  study.  Correlation  of  coefficient  can  be   computed   to   determine   the   predictive   validity   of   the   admission   test.   Predictive   validity   is   strong   if   correlation   of   coefficient   is   greater  than  0.5   o Construct  Validity:  Is  a  conceptual  equation  that  is  developed  by   the   researcher   based   on   theoretical   reasoning.   The   instrument   may  be  considered  to  have  construct  validity  only  if  the  expected   relationships  (between  variable  under  study  and  other  variables)   are  found  to  be  true   • Reliability   • Accuracy/Precision   • Simplicity   • Predictability   Research Problem • Sources   of   Choosing   a   Problem:   Review   of   literature,   academic   experience,   daily   experience,   exposure   to   field   situations,   consultations,   Brain  storming,  Research  and  Intuition   • Formulation  of  problem:   o Internal   Criteria:   Researcher’s   interest,   Researchers   competence   and  Researcher’s  own  resource   o External  Criteria:  Research  ability  of  the  problem,  Importance  and   urgency,  Novelty  of  the  problem,  Feasibility,  Facilities,  Usefulness   &  social  relevance  and  Research  personnel   • Criteria   for   good   research   problem:   Verifiable   evidence,   Accuracy,   precision,   systematization,   objectivity,   recording,   controlling   conditions   and  training  investigators   Hypothesis • Tentative  statement/assumption  asserting  a  relationship  between  certain   facts   • Its  intended  to  be  tested,  verified  or  rejected   • It   contains   variables   that   are   measurable   and   specifying   how   they   are   related   • Criteria   o Not  a  form  of  a  question   o Empirically  testable   o Specific  and  Precise   o Shouldn’t  be  contradictory  
  • Paper1: Research Methodology Exam Sheet   Shivananda  R  Koteshwar,  PhD  Research  Scholar,  Bangalore  University   9   o Should  specify  variables  between  which  the  relationship  is  to  be   established   o Should  describe  only  one  relationship   • Nature  of  Hypothesis   o Accurately  reflect  the  relevant  sociological  fact   o Not  be  in  contradiction  with  approved  relevant  statements  of   other  scientific  disciplines   o Must  consider  the  experience  of  other  researchers   • Characteristics  of  Good  Hypothesis   o Conceptual  Clarity   o Specificity   o Testability   o Availability  of  techniques   o Theoretical  relevance   o Consistency   o Objectivity   o Simplicity   • Types:     o Null  Hypothesis  (H0)  If  we  are  to  compare  method  A  with  method   B  about  its  superiority  and  if  we  proceed  on  the  assumption  that   both  methods  are  equally  good,  then  this  situations  is  termed  as   null   hypothesis.   E.g.   If   we   want   to   test   the   hypothesis   that   the   population   mean   is   equal   to   the   hypothesis   mean   equal   to   100.   Then  null  hypothesis  would  be  H0  :µ=µ  H0  =  100   o Alternative  Hypothesis  (Ha)  If  our  sample  results  do  not  support   this   null   hypothesis,   we   should   conclude   that   something   else   is   true.  What  we  conclude  rejecting  the  null  hypothesis  is  known  as   alternative   hypothesis.   E.g.   For   the   same   example,   the   alternate   hypothesis  are:    Ha:  µ≠µ  H0    -­‐  Population  mean  is  not  equal  to  100    Ha:  µ>µ  H0    -­‐  Population  mean  is  greater  than  100    Ha:  µ<µ  H0    -­‐  Population  mean  is  lesser  than  100   • Level   of   Significance:   If   we   take   level   of   significance   as   5%,   then   this   implies  that  researcher  is  willing  to  take  as  much  as  5%  risk  rejecting  the   null  hypothesis  when  it  happens  to  be  true   • Decision   Rule   of   Test   of   Hypothesis:  Making   rule,   which   is   known   as   decision   rule   according   to   which   we   accept   Null   hypothesis   (rejecting   alternative   hypothesis)   or   reject   null   hypothesis   (accepting   alternative   hypothesis).  E.g.  If  Null  hypothesis  states  that  a  certain  lot  is  good  (less   defective  items)  and  alternate  hypothesis  is  that  the  lot  is  not  good  (many   defective  items).  In  this  case,  we  need  to  decide  the  number  of  items  to  be   tested   and   the   criterion   for   accepting   or   rejecting   the   hypotheses.   We   might  test  10  items  in  the  lot  and  plan  our  decision  saying  that  if  there  are   none   or   only   1   defective   item   among   the   10,   then   we   will   accept   Null   hypothesis   else   we   will   reject   Null   Hypothesis   (and   accept   alternative   hypothesis).  This  sort  of  basis  is  known  as  decision  rule   • Type   1   and   Type   2   Errors   (Type   1   error   is   also   called   as   level   of   significance  of  test)     DECISION  
  • Paper1: Research Methodology Exam Sheet   Shivananda  R  Koteshwar,  PhD  Research  Scholar,  Bangalore  University   10     Accept  NULL   Reject  NULL   Null  Hypothesis  (TRUE)   Correct  Decision   Type  1  Error  (α  error)   Null  Hypothesis  (FALSE)   Type  II  Error  (β  error)   Correct  Decision     • Two  Tailed  Test  and  One  Tailed  Test:     o Two   tailed   test   rejects   the   Null   hypothesis   if,   we   say,   the   sample   mean   is   significantly   higher   or   lower   than   the   hypothesized  value  of  the  mean  of  the  population   o One  tailed  test:  When  we  have  to  say  the  population  mean  is   either  lower  than  or  higher  than  some  hypothesized  value     • Testing  Hypothesis:   o Make   a   formal   statement   -­‐   State   NULL   hypothesis   as   well   as   ALTERNATIVE  hypothesis   o Specify  the  level  of  significance   o Decide  the  correct  sampling  distribution   o Decide  the  sampling  distribution  to  use   o Sample  a  random  sample  and  workout  an  appropriate  value   o Calculate  the  probability  that  sample  result  would  diverge  as   widely   as   it   has   from   expectations,   if   NULL   hypothesis   were   true   o Compare  the  probability  -­‐  If  the  probability  equal  to  or  smaller   than   the   Alpha   value   in   case   of   one   tailed   test   or   equal   to   Alpha/2  in  case  of  two-­‐tailed  test,  reject  NULL  hypothesis  else   accept  NULL  hypothesis   • Tests  of  Significance  or  Tests  of  Hypothesis:   o Parametric  Tests  (Standard  Tests)  –  Assume  certain  properties   of   the   parent   population   from   which   we   draw   samples.   E.g.   sample  size,  population  parameters  like  mean,  variants  etc.  All   tests  are  based  on  the  assumption  of  normality  (Source  of  data   is  considered  to  be  normally  distributed)   o Non  Parametric  Test  or  Distribution  (Free  test  of  hypothesis)  –   Statistical  method   o Important  Parametric  tests    z-­Test:     Used   generally   for   comparing   the   mean   of   a   sample  to  some  hypothesis  mean  for  the  population  in   case   of   large   sample,   or   when   population   variance   is   known.  Based  on  normal  probability  distribution  and  is   used   to   judging   the   significance   of   several   statistical   measures,   particularly   the   mean.   Test   is   also   used   for   both  binomial  distribution  and  t-­‐distribution.    t-­test:   Used   in   case   of   small   sample   when   population   variance   is   unknown.   Based   on   t-­‐distribution   and   is   considered   an   appropriate   test   for   judging   the   significance  of  sample  mean  or  for  judging  significance   of   difference   between   the   two   means   of   the   two   samples.    x2   test:   Used   for   comparing   a   sample   variance   to   a   theoretical   population   variance   is   unknown.   Based   on   chi-­‐square  distribution    
  • Paper1: Research Methodology Exam Sheet   Shivananda  R  Koteshwar,  PhD  Research  Scholar,  Bangalore  University   11    f-­test:   Used   to   compare   the   variance   of   the   two   independent   samples.   This   test   is   also   used   in   the   context  of  variance  (ANOVA)  for  judging  the  significance   of  more  than  2  sample  means  at  the  same  time  and  also   for  judging  the  significance  of  multiple  coefficients.  This   is  based  on  f-­‐distribution   Case Study • Case  study  is  a  method  of  exploring  and  analyzing  the  life  of  a  social  unit   or  entity,  be  it  a  person,  a  family,  an  institution  or  a  community   • The   aim   of   case   study   method   is   to   locate   or   identify   the   factors   that   account  for  the  behavior  patterns  of  a  given  unit  and  its  relationship  with   the  environment   • It   depends   upon   the   wit,   commonsense   and   imagination   of   the   person   doing  the  case  study.   • Efforts   should   be   made   to   ascertain   the   reliability   of   life   history   data   through   examining   the   internal   consistency   of   the   material.   A   judicious   combination  of  techniques  of  data  collection  is  a  prerequisite  for  securing   data  that  are  culturally  meaningful  and  scientifically  significant   • In-­‐depth   analysis   of   selected   cases   is   of   particular   value   to   business   research  when  a  complex  set  of  variables  may  be  at  work  in  generating   observed   results   and   intensive   study   is   needed   to   unravel   the   complexities   Sampling • A  part  of  the  population  is  known  as  sample.  The  method  consisting  of  the   selecting   for   study,   a   portion   of   the   universe   with   a   view   to   draw   conclusions  about  the  universe  or  population  is  known  as  sampling   • Census  (Total  Population)    Target  Population  (Whom  we  are  concerned   with)    Sample  Frame  (Criteria  through  which  we  will  be  selecting)     Sample  Unit  (Categories)    Sample  Element    Sample  Size   • Sample  size  depends  on   o Variability  of  population  (standard  deviation)  –Can  be  found  out   by  Pilot  study   o Confidence  attached  to  the  estimate    (Confidence  Interval)   o Allowable  error  or  margin  of  error  (Tolerable  Error)   • Sample  Size   o Determining  Sample  size  in  case  of  continuous  and  interval  scale    n   =   (Z2   (Std   Dev)2   )   /   (e2)   where   Z   =   Value   of   given   confidence  interval,  n  =  sample  size,  Std  Dev  =  Range/6  and   Range=  Max  Value  -­‐1   o For  Dichotomy  questions    n   =   (Z2   (pq   ))   /   (e2)   where   p=probability   of   success   (frequency  of  people  saying  yes)    If  p  is  not  known,  then  n  =  (¼)(Z2    /  (e2)    
  • Paper1: Research Methodology Exam Sheet   Shivananda  R  Koteshwar,  PhD  Research  Scholar,  Bangalore  University   12   • Sample  Techniques   o Non  Probabilistic:  Convenience  (Accidental),  Judgmental  (Expert   Opinion   or   Purposive),   Quota,   Snowball   (Going   through   references)   o Probabilistic:  Simple  Random,  Systematic  Sample  (E.g.:  Every  5th,   11th,  16th  etc),  Stratified  (homogenous),  Cluster  (Heterogeneous)   o Stratified  can  be  either  proportionate  or  disproportionate     o In  scientific  research  only  probabilistic  sampling  technique  need  to   be  used   o Quota  vs.  Stratified    Quota  is  non  probabilistic  and  Stratified  is  probabilistic    Both   are   homogeneous   within   Quota/Strata   and   heterogeneous  across  Quota/Strata    Both  are  2-­‐stage  process.  In  first  step  Quota  and  Stratified   are   same.   Once   its   Quota   or   stratified,   next   step   would   employ   different   methods.   In   Quota   it   would   be   non   probability   method   and   in   Stratified,   it   would   be   probabilistic     • Quota:   Convenience,   Judgmental   or   Snow   ball   sampling   • Stratified:  Simple  Random  or  Systematic  Random     o Cluster  vs.  Stratified    Heterogeneity   within   Cluster   and   Homogenous   across   cluster      Homogeneous   within   Strata   and   heterogeneous   across   Strata   o Multi   stage   sampling:   Cluster      Stratified      Systematic/Simple   Random   Data Preparation Process • Check  Questionnaire:  Edit,  Code,  Transcribe,  Clean   • Statistically  Adjust  data  /  Statistical  Analysis.  The  two  types  are:   o Descriptive  (Data)   o Inferential  (Hypothesis)        
  • Paper1: Research Methodology Exam Sheet   Shivananda  R  Koteshwar,  PhD  Research  Scholar,  Bangalore  University   13     STATISTICS Distribution: o Normal  Distribution   o Frequency  Distribution  (Poisson,  Binomial,  Normal)   • Discrete  Frequency  Distribution   x   f   74   4   83   3   93   8   • Continuous  Frequency  Distribution   x   f   0-­‐10   4   10-­‐20   3   20-­‐30   8     o For  more  Probability  Distribution:   http://en.wikipedia.org/wiki/Probability_distribution Characteristics of a statistical data • Central  Tendency:  Measured  by  statistical  averages   o Mathematical  Average:  Arithmetic  Mean,  Geometric  Mean,   Harmonic  Mean   o Positional  Average:  Median,  Mode   • Dispersion   • Skewness   • Kurtosis   Arithmetic Mean o AM=∑X/N  where  ∑X  =  Sum  of  the  item  and  N  is  the  number  of  items   o If   frequency   is   given,   then   AM=∑fx/∑f   where   ∑fx   =   sum   of   the   values   multiplied  by  the  corresponding  frequency  and  ∑f    is  sum  of  frequency     o Arithmetic  mean  of  58,67,68,84,93,98,100    ∑X    =  58+67+68+84+93+98+100  =  560    N  =  number  of  items  =  7    AM  =  ∑X/N    =  560/7  =  80     o Arithmetic  mean  of  following  50  workers  according  to  their  daily  wages    Daily  Wage:  15,  18,  20,  25,  30,  35,  40,  42,  45    Number  of  workers:  2,  3,  5,  10,  12,  10,  5,  2,  1   Wages   (x)   Frequency  (F)   fx   15   2   30  
  • Paper1: Research Methodology Exam Sheet   Shivananda  R  Koteshwar,  PhD  Research  Scholar,  Bangalore  University   14   18   3   54   20   5   100   25   10   250   30   12   360   35   10   350   40   5   200   42   2   84   45   1   45    ∑fx    =  473  and  ∑f    =  50    AM  =∑fx/∑f    =  473/50  =  29.46     o Arithmetic  mean  for  the  following  distribution    Marks  10-­‐20  20-­‐30  30-­‐40  40-­‐50  50-­‐60  60-­‐70  80-­‐90    Number  of  students:  6  12  18  20  20  14  8  2   Marks   Frequency  (F)   Mid  Value  (x)   Mean   fx   10-­‐20   6   15   90   20-­‐30   12   25   300   30-­‐40   18   35   630   40-­‐50   20   45   900   50-­‐60   20   55   1100   60-­‐70   14   65   910   70-­‐80   8   75   600   40-­‐90   2   85   170      ∑fx    =  4700  and  ∑f    =  100    AM  =  =∑fx/∑f    =  4700/100  =  47   Median • Size  of  the  middlemost  value   • 80,  86,  74,  465,  3,  984,  22:  Median  is  465   • Median  of  Indian  age  is  26  means,  50%  of  India’s  population  will  be  less   than  26years  of  age  and  50%  will  be  more  than  26yrs  of  age   Mode • Most  occurring  number   Standard Deviation and Variance o Deviation  from  Mean   o It’s  a  relative  number  and  not  an  absolute  number   o Lesser  the  Standard  Deviation,  higher  the  reliability   o σ  =  √(∑(x-­‐xb)2  /  N)   x   (x-­‐xb)2    15    64   20   9   22   1   28   25  
  • Paper1: Research Methodology Exam Sheet   Shivananda  R  Koteshwar,  PhD  Research  Scholar,  Bangalore  University   15   30   49   ∑x  =  115     ∑(x  –xb)2  =  148   • xb  =  ∑x/N    =  115/5  =  23   • σ  =  √(148/5)  =  5.44   • Variance  =  σ2  =  29.59     Coefficient  of  Variation   • Lesser  the  confidence  of  variation,  the  reliability  is  higher   • V  =  σ  /xb*100   • For  the  above  example,  it  would  be  equal  to  5.44/23*100  =  23.65   • Lesser  the  CV,  higher  the  reliability   Range  and  Coefficient  of  Range   • Range  =  L-­‐S   • Coefficient  of  Range  =  (L-­‐S)/(L+S)   Trend Analysis (Straight Line Analysis) • Least  Square  Method  (Forecasting  Method)   Year   Sales   (y)   year-­midyear   x   x2   xy   yc   bx  +  a   2006   42   -­‐3.5   12.25   -­‐147   36.11   2007   40   -­‐2.5   6.25   -­‐100   41.97   2008   36   -­‐1.5   2.25   -­‐54   47.83   2009   58   -­‐0.5   0.25   -­‐29   53.69   2010   62   0.5   0.25   31   59.55   2011   60   1.5   2.25   90   65.41   2012   70   2.5   6.25   175   71.27   2013   80   3.5   12.25   280   77.13     ∑y=453   ∑x=0   ∑x2=42   ∑xy  =  246   ∑yc  =452.96     • Mid  year  =  2009.5   • Deviation  from  Arithmetic  mean  will  be  least  in  this  method,  hence  its   called  least  square  method   • yc  =  bx  +  a   • ∑y  =  b∑x  +  Na     • ∑y  =  b  (0)+  Na  =  Na   • a  =  ∑y  /N   • a  =  453/8  =  56.62   • ∑xy  =  a∑x  +  b∑x  2   • 246=  56.62  (0)  +  b  (42)   • b  =  5.86   • Forecast  for  2014,  x  =  4.5   o yc  =  bx  +  a   o yc  =  5.86  (4.5)  +  56.62  =  82.99  
  • Paper1: Research Methodology Exam Sheet   Shivananda  R  Koteshwar,  PhD  Research  Scholar,  Bangalore  University   16   • Forecast  for  2015,  x  =  5.5   o yc  =  bx  +  a   o yc  =  5.86  (5.5)  +  56.62  =  88.85   • ∑yc  =  ∑y    (Verification  Technique)   Standard  Normal  Curve  (SNC)   1. Assume  mean  height  of  soldier  is  68.22  inches  with  a  variance  of  10.8   inches.  How  many  soldiers  in  a  regiment  of  1000  would  you  expect  to   be  over  6ft  tall   • σ  2  =  10.8   • σ  =  3.29   • x  =  6  feet  =  72  inches   • xb  =  68.22  (mean)   • z  =  SNC  =  (x-­‐xb)/σ    =  (72-­‐68.22)/3.29  =  1.15     • From  the  Statistical  Table  for  1.15  its  =>  0.5  –  0.3759  =  0.1251   • 0.1251*1000  =125  soldiers  are  taller  than  1000     2. 1000  light  bulbs  with  a  mean  life  of  120  days  are  installed  in  a  new   factory.  They  have  length  of  life  is  normally  distributed  with  Standard   deviation  of  20  days.  How  many  bulbs  will  expire  in  less  than  90  days?   How  many  bulbs  will  burn  for  more  than  125  days?   • N  =  1000   • xb  =  120   • σ  =  20   • x  =90   • Z  =  SNC  =  (x-­‐xb)/σ  =  (90-­‐120)/20  =  -­‐1.5     • From  the  statistical  table  for  -­‐1.5  its  =>  0.5  -­‐0.4332  =  0.0668   • 0.0668*1000  =  67  Bulbs     • N  =  1000   • xb  =  120   • σ  =  20   • x  =125   • Z  =  SNC  =  (x-­‐xb)/σ  =  (125-­‐120)/20  =  0.25     • From  the  statistical  table  for  0.25  its  =>  0.5  -­‐0.0987  =  0.4013   • 0.4013*1000  =  401  bulbs   Non parametric test – (χ2) kai2 test o χ2     =   kai2     =   ∑   (O-­‐E)2/E2       where   O   =   Observed   Frequency   and   E   =   Expected  frequency   o In   a   certain   area   in   Bangalore,   the   corporation   distributed   pills   to   combat  CG.  From  the  data  given  below  analyze  whether  the  pills  given   were  effective  or  not  in  combating  the  disease  
  • Paper1: Research Methodology Exam Sheet   Shivananda  R  Koteshwar,  PhD  Research  Scholar,  Bangalore  University   17     Fell  Ill   Not  Ill   Took  Pills   345   620   Dint  take   pills   545   450     o Null  Hypothesis:  Given  pills  are  not  effective  in  controlling  the  said   disease     Table  of  Observed  Frequency  (O)   345   620   965  (Row  1  Total)   545   450   995  (Row  2  Total)   890  (Column1  Total)   1070  (Column  1   Total)   1960  (Grand   Total)     • E  =  (RT  *  CT)  /  GT     • Table  of  Expected  Frequency  (E)     • E345  =  965*890  /  1960  =  438.19   • E620  =  965*1070  /  1960  =  526.81   • E545  =  995*890  /  1960  =  451.81   • E450  =  995*1070/  1960  =  543.19     438.19   526.81   965     451.81   543.19   995     890     1070     1960       O   E   (O-­E)2/E2   345   438.19   0.045   545     451.81   0.042   620   526.81   0.032   450   543.19   0.029       ∑  (O-­‐E)2/E2      =   0.148     • χ2  =  Kai2  =  0.148   • Degree  of  freedom    =  (r-­‐1)  (c-­‐1)  =  (2-­‐1)(2-­‐1)  =  1   • Taking  the  significance  level  to  be  5%  (Confidence  level  as  95%),   from  the  statistical  table,  we  can  find  that  the  table  value  is  3.84   • As   calculated   hypothesis   =   0.1484   is   less   than   the   table   value   of   3.84,  Null  hypothesis  is  accepted   ANNOVA – Analysis of Variance 1. 5   salesmen   work   in   4   cities.   Based   on   the   data   given   determine   whether  there  is  a  significance  difference  in  the  sales  performance  of   different  cities   Salesmen   A   B   C   D   S1   14   12   13   15   S2   15   14   12   11  
  • Paper1: Research Methodology Exam Sheet   Shivananda  R  Koteshwar,  PhD  Research  Scholar,  Bangalore  University   18   S3   16   17   15   10   S4   12   16   15   14   S5   10   11   15   17     • Null  Hypothesis:  There  is  no  significance  difference  in  the  sale   performance  of  different  cities     X1   X2   X3   X4     14   12   13   15     15   14   12   11     16   17   15   10     12   16   15   14     10   11   15   17   ∑X   67   70   70   67   Xb  =  ∑X/N  (N=5)   13.4   14   14   13.4     • Grand  Mean  =  Xbb  =  ∑Xb/N  =  (13.4  +  14  +  14  +  13.4)/4  =  13.7     • Variance  between  samples     (X1b-­ X1bb)2   (X2b-­ X2bb)2   (X3b-­ X3bb)2   (X4b-­X4bb)2     0.09   0.09   0.09   0.09     0.09   0.09   0.09   0.09     0.09   0.09   0.09   0.09     0.09   0.09   0.09   0.09     0.09   0.09   0.09   0.09   ∑   0.45   0.45   0.45   0.45     • Sum  of  Squares  =  0.45  +  0.45  +  0.45  +  0.45  =  1.8   • Degree  of  Freedom  (d.f)  γ1    =  4  -­‐1  =  3     • Mean  of  sum  of  squares  =  1.8/3  =  0.6     • Variance  within  samples     (X1-­X1b)2   (X2-­X2b)2   (X3-­X3b)2   (X4-­X4b)2     0.36   4   1   2.56     2.56   0   4   5.76     6.76   9   1   11.56     1.96   4   1   0.36     11.56   9   1   12.96   ∑   23.2   26   1   33.2     • Sum  of  Squares  =  23.2  +  26  +  1  +  33.2  =  90.4   • Degree  of  Freedom  (d.f)  γ2  =  Total  number  of  observations  –  Number   of  samples  =  (5*4)  –  4  =  16     • Mean  of  sum  of  squares  =  90.4/16  =  5.65   •   • “f”  test  (Fisher)  for  5%  significance  level   • f  test  =  F=  (variation  between  samples)/(variation  within  samples)  =   0.6/5.65  =  0.106     • From  the  table,  m=γ1  and  n=γ2  ,  m=3  and  n=16,  value  of  F=3.2389  
  • Paper1: Research Methodology Exam Sheet   Shivananda  R  Koteshwar,  PhD  Research  Scholar,  Bangalore  University   19     • Calculated  value  F=0.106  is  less  than  the  table  value  3.2389  so  Null   Hypothesis  is  accepted     2. 5   salesmen   work   in   4   cities.   Based   on   the   data   given   determine   whether   there   is   a   significance   difference   between   the   sales   performance  of  different  salesmen   Salesmen   A   B   C   D   S1   14   12   13   15   S2   15   14   12   11   S3   16   17   15   10   S4   12   16   15   14   S5   10   11   15   17         S1   S2   S3   S4   S5   City1   14   15   16   12   10   City2   12   14   17   16   11   City3   13   12   15   15   15   City4   15   11   10   14   17   ∑X   54   52   58   57   53   Xb=∑X/N  (N=4)   13.5   13   14.5   14.25   13.25     • Grand  Mean  =  Xbb  =  ∑Xb/N  =  (13.5  +  13  +  14.5  +  14.25  +  13.25)/5=   13.7     • Variance  between  samples     (X1b-­X1bb)2   (X2b-­X2bb)2   (X3b-­ X3bb)2   (X4b-­ X4bb)2   (X5b-­ X5bb)2     0.04   0.49   0.64   0.3   0.2     0.04   0.49   0.64   0.3   0.2     0.04   0.49   0.64   0.3   0.2     0.04   0.49   0.64   0.3   0.2   ∑   0.16   1.96   2.56   1.2   0.8     • Sum  of  Squares  =  0.16  +  1.96  +  2.56  +  1.2  +  0.8    =  6.68   • Degree  of  Freedom  (d.f)  γ1    =  5  -­‐1  =  4     • Mean  of  sum  of  squares  =  6.68/4  =  1.67     • Variance  within  samples     (X1-­X1b)2   (X2-­X2b)2   (X3-­X3b)2   (X4-­X4b)2   (X5-­X5b)2     0.25   4   2.25   5.06   10.56     2.25   1   6.25   3.06   5.06     0.25   1   0.25   0.56   3.06     2.25   4   20.25   0.06   14.06   ∑   5   10   29   8.74   32.74     • Sum  of  Squares  =  5  +  10  +  29  +  8.74  +  32.74  =  85.48   • Degree  of  Freedom  (d.f)  γ2  =  Total  number  of  observations  –  Number   of  samples  =  (5*4)  –  5  =  15    
  • Paper1: Research Methodology Exam Sheet   Shivananda  R  Koteshwar,  PhD  Research  Scholar,  Bangalore  University   20   • Mean  of  sum  of  squares  =  85.48/15  =  5.7   •   • “f”  test  (Fisher)  for  5%  significance  level   • f  test  =  F=  (variation  between  samples)/(variation  within  samples)  =   1.67/5.7  =  0.29     • From  the  table,  m=γ1  and  n=γ2  ,  m=4  and  n=15,  value  of  F=3.0556     • Calculated  value  F=0.29  is  less  than  the  table  value  3.0556  so  Null   Hypothesis  is  accepted   Coefficient of Correlation • Carls  Coefficient  Method   • r  =    ∑xy  /  (√(∑x2  *  ∑y2)  where  x  =  X-­‐Xb  and  y  =  Y-­‐Yb     • Calculate  the  coefficient  of  correlation  for  the  following  value   X   Y   24   16   36   22   32   34   38   48   40   60       X   Y   x=X-­‐Xb     y=Y-­‐Yb     x2   y2   xy     24   16   -­‐10   -­‐20   100   400   200     36   22   +2   -­‐14   4   196   -­‐28     32   34   -­‐2   -­‐2   4   4   4     38   48   4   4   16   144   48     40   60   6   6   36   576   144   ∑   170   180       160   1320   368     • Xb  =  ∑X/N  =  170/5  =  34   • Yb  =  ∑Y/N  =  180/5  =  36   • r  =    ∑xy  /  (√(∑x2  *  ∑y2)    =  368/(√(160*1320)  =  0.8   Regression o x  on  y    X-­‐Xb=  bxy  (Y-­‐Yb)    bxy  =  ∑xy  /  ∑y2     o y  on  x    Y-­‐Yb=  byx  (X-­‐Xb)     byx  =  ∑xy  /  ∑x2     o Calculate  the  regression  for  the  following  table     X   Y   32   12   48   15   24   18  
  • Paper1: Research Methodology Exam Sheet   Shivananda  R  Koteshwar,  PhD  Research  Scholar,  Bangalore  University   21   26   25   30   20       X   Y   x=X-­‐Xb     y=Y-­‐Yb     x2   y2   xy     32   12   -­‐0   -­‐6   0   36   0     48   15   16   -­‐3   256   9   -­‐48     24   18   -­‐8   0   64   0   0     26   25   -­‐6   7   36   49   -­‐42     30   20   -­‐2   2   4   4   -­‐4   ∑   160   90       360   98   -­‐94     • Xb  =  ∑X/N  =  160/5  =  32   • Yb  =  ∑Y/N  =  90/5  =  18     • x  on  y   • bxy  =  ∑xy  /  ∑y2    =  -­‐94/  98  =  -­‐  0.96   • X-­‐Xb=  bxy  (Y-­‐Yb)   • X-­‐32  =  -­‐0.96(Y-­‐18)  =>  X=  -­‐0.96Y  +  49.28     • y  on  x   • byx  =  ∑xy  /  ∑x2    =  -­‐94/360  =  -­‐0.26   • Y-­‐Yb=  byx  (X-­‐Xb)   • Y-­‐18  =  -­‐0.26(X-­‐32)  =>  Y  =  -­‐0.26X  +  26.32   Small Sample Test o T  Test  (Student  Test)  when  sample  size  is  less  than  30   o t  =  (Xb  -­‐  µ0)  /  (s/√n-­‐1)     1. The  mean  percentage  of  passes  in  all  the  schools  of  a  town  was  found   to   be   83%.   A   random   sample   of   17   schools   revealed   that   86%   pass   with  standard  deviation  of  3%.  Test  a  1%  level  of  significance  whether   the  mean  percentage  of  passes  is  more  than  83%   • µ  =  83%   • n  =  17   • Xb  =  86%   • s=  3%   • Degree  of  freedom  =  n-­‐1  =  16   • Level  of  Significance  =  1%     • Null  Hypothesis  (H0):  Mean  percentage  of  passes  is  less  than  83%   • Alternate  Hypothesis  (Ha):  Mean  percentage  of  passes  is  more  than   83%     • tcal  =  (Xb  -­‐  µ0)  /  (s/√n-­‐1)  =  (86-­‐83)  /  3/√(17-­‐1)  =  4     • From  the  statistics  table,  ttable  =  2.583  (For  Degree  of  freedom  =  16  and   Level  of  significance  of  1%)    
  • Paper1: Research Methodology Exam Sheet   Shivananda  R  Koteshwar,  PhD  Research  Scholar,  Bangalore  University   22   • ttable  <  tcal  implies  that  the  Null  Hypothesis  is  in  critical  region  so  its  not   accepted  so  Ha  is  accepted      
  • Paper1: Research Methodology Exam Sheet   Shivananda  R  Koteshwar,  PhD  Research  Scholar,  Bangalore  University   23     IMPORTANT QUESTIONS 1. Distinguish   between   probability   and   non   probability   sampling   methods’  by  giving  suitable  examples   2. Research  refers  to  ends  and/or  means.  Discuss  this  statement   3. Hypothesis  is  the  guiding  force  in  any  research  study?  Justify  and   explain  the  process  of  hypothesis  formulation  and  testing  it  with   suitable  example   4. Briefly  describe  the  contents  of  a  research  report   5. Briefly  describe  the  various  methods  used  for  descriptive  analysis   of  data   6. What   is   sampling?   List   the   similarities   and   differences   between   stratified  sampling  and  quota  sampling   7. How   are   research   design   classified?   What   are   the   distinguishing   features  of  each?  Differentiate  by  giving  appropriate  examples   8. What   do   you   mean   by   measurement?   Explain   four   key   levels   of   measurement  with  suitable  examples  and  also  give  details  of  what   statistical  technique  can  be  used  with  data  from  each  type  of  scale?   9. What   is   scaling?   Describe   the   various   comparative   and   non   comparative   scaling   techniques   used   in   business   research   with   suitable  examples   10. How  do  you  edit  a  questionnaire?  What  are  the  precautions  that  a   research  must  take  while  editing  and  coding  a  questionnaire?  Give   suitable  example   11. Explain   various   Parametric   and   Non   Parametric   Test   with   examples   12. Discus  the  various  types  of  research  and  their  features   13. Find   the   correlation   of   coefficient   for   the   following   data   and   comment  on  its  significance?     X   24   26   36   35   43   45   47   Y   47   48   54   58   59   59   65     14. Perform   ANOVA   with   5%   level   of   significance   to   determine   whether  there  is  a  significant  difference  in  the  mean  speed  of  four   different  machines   Hours   Machine  A   Machine  B   Machine  C   Machine  D   1   15   14   30   35   2   20   16   25   30   3   25   22   24   32   4   20   28   26   28     15. From   the   data   given   below   about   the   treatment   of   patients   suffering  from  cold,  state  whether  the  new  treatment  is  superior  to   that   of   the   conventional   treatment.   You   can   use   Kai2   test   for   evaluation   Treatment   Favorable   Not  Favorable   New   280   60   Conventional   120   40  
  • Paper1: Research Methodology Exam Sheet   Shivananda  R  Koteshwar,  PhD  Research  Scholar,  Bangalore  University   24     16. Calculate  the  straight  line  trend  for  the  following  data  and  forecast   the  production  figures  for  the  next  two  years   Year   2006   2007   2008   2009   2010   2011   2012   2013   Production   43   67   34   76   71   85   88   96     17. A  cooperative  wishes  to  test  whether  the  preference  of  consumers   for  its  products  its  dependent  on  income  levels.  Use  the  Chi  square   test  to  decide.  You  may  use  a  5%  significance  level   Product  Preferred   Income   Product  A   Product  B   Product  C   Product  D   Low   185   45   95   325   Medium   65   40   75   180   High   35   25   70   130   Total   285   110   240   635