Banner Ad Location Effectiveness Study: Statistics | WebReference

Banner Ad Location Effectiveness Study: Statistics

Statistical Significance Calculations

Banner Ad Placement Study

Hypotheses Testing

The following hypothesis test was performed for each set of results to determine if the difference due to location of the banner ad was significant.

Null Hypothesis: p1-p2=0

Test Statistic: z=(p1-p2)/sqrt(p*q(1/m +1/n))

p1= clicks/banner impressions sample 1
p2= clicks/banner impressions sample 2
m= banner impressions sample 1
n= banner impressions sample 2
p=m/(m+n) * p1 + n/(n+m) * p2
q= 1-p

P-Value: The smallest level of significance at which the null hypothesis would be rejected. (In other words, it is the smallest level at which the data is significant.) A two-tailed P-value was calculated for this data. P-values less than 0.05 were considered to be statistically significant.

Sample
m
n
p1
p2
p
z
Norm Dist
2 sided p-value
Statistically Significant?
Star Wars 1
6972
6025
0.014
0.034
0.024
-7.318
1.000
0.000
Yes
Star Wars 2
4895
5830
0.007
0.018
0.013
-5.013
1.000
0.000
Yes
Digital Frontiers 1
321
299
0.028
0.134
0.079
-4.886
1.000
0.000
Yes
Digital Frontiers 2
366
400
0.052
0.173
0.115
-5.240
1.000
0.000
Yes
Webreference
1683
1593
0.011
0.014
0.012
-0.774
0.780
0.439
No
Photodisc
1423
2027
0.022
0.039
0.032
-2.793
0.997
0.005
Yes
HoloNet
2374
2352
0.005
0.004
0.004
0.258
0.602
0.796
No

Back to Home

Comments are welcome


Created: Apr. 21, 1997
Revised: Apr. 27, 1997

URL: http://webreference.com/dev/banner/statistics.html