International Pre-Masters Diploma in Business Studies.International Pre-Masters Diploma in Legal Studies.Project & Report Session: Quantitative & Qualitative data Liam Greenslade Professional Education: developing your career
Overview / Learning outcomes Students should be able to answer these questions: What are the key differences between quantitative and qualitative approaches? What is meant by hypothesis testing, and why does it require quantification? Why do researchers use quantitative methods? What is meant by a level of measurement? Why does the level of measurement matter?
Qualitative & Quantitative data Quantitative data is made up of numbers, measurements & statistics It is mainly collected by positivists who try to study in a systematic and objective way and use logic and the scientific approach Qualitative data is collected by interpretivists who are concerned with meaning and the interpretation of experience
Qualitative vs Quantitative Data Qualitative Data Quantitative DataOverview: Overview:•Deals with descriptions. •Deals with numbers.•Data can be observed but •Data which can benot measured. measured.•Colors, textures, smells, •Length, height, area, volume,tastes, appearance, beauty, weight, speed, time,etc. temperature, humidity, sound levels, cost, members, ages,•Qualitative → Quality etc. •Quantitative → Quantity
Example 1: Oil Painting Example 1: Oil PaintingQualitative data: Quantitative data:*red/orange color, no frame *picture is 100" by 24”*smells old and musty * weighs 8.5 pounds*texture shows brush strokes of oil paint •surface area of painting is 2400*winter sunset scene sq. in.*masterful brush strokes *cost £300
Example 2: Cappuccino Example 2: CappuccinoQualitative data: Quantitative data: *robust aroma *12 ounces of coffee *frothy appearance *serving temperature 1500 F. * strong taste *serving cup 7 inches in height *glass cup *cost £2.95
Qualitative vs. Quantitative Research Qualitative Quantitative Research ResearchPurpose Discover ideas Test hypotheses or specific research questionsApproach Observe and interpret Measure and testData Collection Unstructured; free- Structured; responseMethods forms categories providedResearcher Researcher is Researcher isIndependence intimately involved; uninvolved; results are results are subjective objectiveSample Small samples – often Large samples to allow natural setting generalizationMost often used in: Exploratory research Descriptive and causal designs research designs
Qualitative vs. Quantitative Research Qualitative Research Quantitative ResearchObjective To gain a qualitative To quantify the data and understanding of the generalize the results from underlying reasons and the sample to the population motivations of interestSample Small number of non- Large number of representative cases representative casesData Collection Unstructured StructuredData Analysis Non-statistical StatisticalOutcome Develop an initial Recommend a final course of understanding action
Research Methods Type Of Method Theoretical Data PerspectivePrimary quantitative Experiments Positivist qualitative Observation Interpretivist quantitative Questionnaires Structured Positivist interviews qualitative Unstructured Interpretivist Interviews Positivist quantitativeSecondary Official Statistics qualitative Mass Media, Interpretivist
Why do some researchers chooseto use qualitative data? Gives priority to meaning and experience Provides richer and deeper understanding of human phenomena Qualitative data can be very powerful in assessing feelings about issues
Why do some researchers chooseto use quantitative data? Accurate answers to empirical questions may require numerical data. Common sense perceptions are often inaccurate Some research employs hypotheses testing. Hypothesis testing is based on precise, testable predictions. Quantification allows a precise prediction, against which an outcome can be tested. Predictions can be tested using statistical analysis.
Logic of hypothesis testing Hypothesis testing is more associated with natural/physical sciences, but also employed by researchers in business Related to the idea that a scientific theory should allow predictions Predictions are tested against data To be testable, a theory must exclude certain things from happening; i.e. there must be a possible outcome which shows that the theory is wrong
Hypothesis testing & interpretativemethods Hypothesis testing : making & testing predictions about events or behaviour from theory and past findings (top-down) Associated with use of quantifiable data Interpretative methods : analysis of meaning in data; gaining insight into participant’s experiences (bottom up) Associated with use of qualitative data
Using quantitative data Quantitative data is used in various ways in social research: To report differences between groups in some numerical measurement (e.g. males versus females, high versus low socioeconomic areas) To report associations between different sets of scores or measurements (e.g. associations between income and happiness, or crime rates and income). The measurable effects can be used to describe social phenomena, and/or to test predictions
Cause & effect Where the research involves testing hypotheses about causal relationships, the terms ‘independent’ and ‘dependent’ variable are used Independent variable (IV): the presumed cause; the variable that produces an effect Dependent variable (DV): the outcome; the variable that is affected.
Concept ofYCausalityA statement such as "X causes " will have thefollowing meaning to an ordinary person and to ascientist.____________________________________________________ Ordinary Meaning Scientific Meaning____________________________________________________X is the only cause of Y. X is only one of a number of possible causes of Y.X must always lead to Y The occurrence of X makes the(X is a deterministic occurrence of Y more probablecause of Y). (X is a probabilistic cause of Y).It is possible to prove We can never prove that X is athat X is a cause of Y. cause of Y. At best, we can infer that X is a cause of Y.
Conditions for Causality Concomitant variation is the extent to which a cause, X, and an effect, Y, occur together or vary together in the way predicted by the hypothesis under consideration. The time order of occurrence condition states that the causing event must occur either before or simultaneously with the effect; it cannot occur afterwards. The absence of other possible causal factors means that the factor or variable being investigated should be the only possible causal explanation.
Experimental & non-experimentalquantitative approaches Experimental approaches aim to establish cause and effect relationships between variables The researcher manipulates the IV to determine its effects on the dependent variable When the variable treated as an IV cannot be manipulated, this is termed a quasi-experiment Non-experimental approaches examine existing relationships between variables rather than attempting to produce effects
What is a variable? A variable = something that can vary! • More specifically, a variable has a range of possible numerical values • Quantification uses the number system to represent properties of objects or events Variables in social & behavioural sciences are often not easy to measure (e.g. socio-economic status, intelligence, crime rates,, attitude).
Nature of research variables Can be quantified (represented by numbers) Free to vary (take on any of a range of possible numerical values) Distinguishable from each other (so we can quantify the right one).
How do we make variablesmeasurable? Operational definition – defining a variable in terms that make clear how it will be measured This is described as ‘operationalising’ the dependent variable This is essential for using the variable in quantitative research Social scientists often work with variables which are difficult to operationalise.
Levels of measurement Nominal – observations can be assigned to a category (a distinct group). There is no logical order for categories. Numbers act like names Ordinal – observations can be assigned values that are rank ordered (arranged in a logical order), where one rank is further along a dimension than another (e.g. age groupings) Interval – measures can be taken on a continuous scale where equal intervals represent equal differences (e.g. Thermometer) Ratio – measures are taken on a continuous scale with interval properties and a true zero point (e.g. age, length, height, weight)
Nominal, Ordinal, Interval, and Ratio Scales Provide Different Information
Why do levels of measurement matter? Numbers represent different things and have different functions at each level of measurement Nominal : we can classify or categorise observations. Numbers represent frequency counts. Ordinal : we can rank or order observations. Numbers represent rank or relative position in a sequence. Interval/ratio : we can measure characteristics or properties and assign them scores. Scores represent an exact amount of some property.
How do levels of measurementaffect analysis? Levels of measurement determine what can be done with data Nominal data can only be used to take frequency counts (how many observations fall into each category) Ordinal data can be arranged from highest to lowest, but statistical analyses are limited Interval/ratio – we know exactly how much scores differ from each other, can calculate the mean (average) score and how much scores vary from each other; can use more statistical tests
Working with measured variables When we have scores at an interval level, we can do much more with data Measure the central tendency (e.g. Calculate the mean) Measure the variability (e.g. Calculate the standard deviation). These measures are important in statistical analysis and will be covered in later sessions.
Ordinal scales treated as interval Some measures only appear interval, for example: ‘Harsher penalties would reduce violent crime’ Agree Disagree 1 2 3 4 5 6 We cannot be sure that this ‘Likert scale’ has equal distances between intervals. In practice though, it is usually treated as interval.
How is an IV manipulated? – True experiment A researcher wants to test the effects of frustration on aggression. Participants are randomly assigned to an experimental group exposed to frustration, and a control group which is not A measure of aggression is taken and the mean aggression scores of the experimental and control group are compared
When can an IV not be manipulate? –Quasi-experiment A researcher wants to know if some personality types are more likely to become aggressive when frustrated Participants are assigned to groups based on personality type. However, this assignment cannot be random because personality type is predetermined. The participants are exposed to frustration and the mean aggression scores of the personality types are compared This is a quasi-experiment because the researcher tried to produce an effect (aggression), but the variable treated as the IV (personality) cannot be manipulated
Examples of non- experimental quantitative methods Surveys Content analysis Naturalistic observation Participant observation Archival research / secondary analysis All of these methods examine effects that have already occurred rather than effects produced by the researcher. Analysis often looks at correlation (whether changes in one variable are associated with changes in another). Because there is no experimental control, there are difficulties inferring which variables are causing the effects observed. However, more variables can be measured at once and the phenomena observed occur naturally.