This study examined the relationship between mobile device attributes and user-perceived quality of Android apps. The researchers analyzed 150,373 star ratings from Google Play across 30 devices and 280 apps. They found that the perceived quality of apps varies across devices, and having better characteristics of an attribute does not necessarily correlate with higher quality. Device OS version, resolution, and CPU showed significant relationships with ratings, as did some app attributes like lines of code and number of inputs. However, some device attributes had stronger relationships than app attributes.

1. A study of the relation of mobile device
attributes with the user-perceived quality of
Android apps
1
Ehsan Noei, Mark D. Syer, Ying Zou, Ahmed E. Hassan, and Iman
Keivanloo

2. Different mobile devices are available in the market.
2

3. The number of mobile apps continues to rise.
3
0
0.5
1
1.5
2
2.5
3
3.5
4
Dec
'09
Apr
'10
Oct
'10
Jul
'11
Dec
'11
May
'12
Sep
'12
Apr
'13
Jul
'14
Jul
'15
Feb
'16
Dec
'16
Jun
'17
Dec
'17
Month
MillionApps

4. 4
Stupid App!

5. Users rely on star-ratings when choosing a new app.
5

6. Do devices have a relationship with
the user-perceived quality of Android
apps?
Which device attributes share the
most significant relationships with the
user-perceived quality?
Do device attributes have a stronger
relationship with the user-perceived
quality than app attributes?
RQ1)
RQ2)
RQ3)
6

7. 7
Google Play Store
Data Collection

8. 8
Google Play Store
User-reviews
APK Files

9. 9
Google Play Store
User-reviews
APK Files

10. 10
Google Play Store
150,373
star-ratings
30 mobile devices
Bytecode of 280 Apps

11. RQ1) Do devices have a relationship with the
user-perceived quality of Android apps?
11

12. The user-perceived quality of apps varies across
devices.
12

13. The user-perceived quality of apps varies across
devices.
13

14. RQ2) Which device attributes share the most
significant relationships with the user-perceived
quality?
14

15. Mixed effects model
Dependent variable
Independent variables
Mixed effect variables
Star-ratings
Device attributes
App and device
15

16. Display Size
Resolution
OS Version
CPU
GPU
Internal RAM
Internal ROM
Camera
Talk Time
etc.
16
Device Attributes

17. Mixed effects model
Dependent variable
Independent variables
Mixed effect variables
Star-ratings
Device attributes
App and device
17

18. 18
OS Version
Display
Resolution
CPU
Standby
Time
Camera GPU Battery Device Size

19. Device attributes share a significant
relationship with the user-perceived quality.
19
OS Version
Display
Resolution
CPU
Standby
Time

20. Having a better characteristic of an attribute
does not necessarily share a positive
relationship with the user-perceived quality of
apps.
20
↘ ↘ ↗ ↗
OS Version
Display
Resolution
CPU
Standby
Time

21. RQ3) Do device attributes have a stronger
relationship with the user-perceived quality than app
attributes?
21

22. Mixed effects model
Dependent variable
Independent variables
Mixed effect variables
Star-ratings
Device attributes + App
attributes
App and device
22

23. Mixed effects model
Dependent variable
Independent variables
Mixed effect variables
Star-ratings
Device attributes + App
attributes
App and device
23

24. Display Size
Resolution
OS Version
CPU
GPU
Internal RAM
Internal ROM
Camera
Talk Time
etc.
24
LOC
#Methods
WMC
Installation Package Size
UI Depth
#UI Nodes
#Inputs
#Outputs
#Activities
etc.
App AttributesDevice Attributes

25. Mixed effects model
Dependent variable
Independent variables
Mixed effect variables
Star-ratings
Device attributes + App
attributes
App and device
25

26. 26
OS Version
Display
Resolution
CPU
Standby
Time

27. 27
OS Version
Display
Resolution
CPU
Standby
Time
LOC WMC
#Inputs

28. 28
Both the app and the device attributes have
significant relationships with the user-perceived
quality.
OS Version
Display
Resolution
CPU
Standby
Time
LOC WMC
#Inputs

29. Attribute Effect
LOC *** ↘
WMC *** ↘
#Inputs * ↗
APK Size ↗
OS Version *** ↘
Resolution ** ↘
CPU * ↗
Standby Time * ↗
Photo Size (MP) + ↘
Internal ROM + ↘
29
Some device attributes have stronger relationships
with the user-perceived quality than some app
attributes.

30. 30

31. The user-perceived quality of apps varies across
devices.
31

32. 32

33. Having a better characteristic of an attribute
does not necessarily share a positive
relationship with the user-perceived quality of
apps.
33
OS Version
Display
Resolution
CPU Standby Time
↘ ↘ ↗ ↗

34. 34

35. 35
Both the app and the device attributes have
significant relationships with the user-perceived
quality.
OS Version
Display
Resolution
CPU
Standby
Time
LOC WMC
#Inputs

36. 36

37. Attribute Effect
LOC *** ↘
WMC *** ↘
#Inputs * ↗
APK Size ↗
OS Version *** ↘
Resolution ** ↘
CPU * ↗
Standby Time * ↗
Photo Size (MP) + ↘
Internal ROM + ↘
37
Some device attributes have stronger relationships
with the user-perceived quality than some app
attributes.

38. 38