One-sample t-test

Ho: u = 5 (status quo bias) Ha: = >5 (one tail test) If you put a certain statement as your null hypothesis, you are making it difficult to be rejected. Common Lqaw: Ho: the suspect is not quilty If p-value >= 0.05, do not reject Ho; otherwise reject Ho. Since p-value (SIG) = 0.367/2 = 0.1835 > 0.05, do not reject Ho. The true buying intention is not significantly different from 5, so we should not invest money into Weelworth ** SPSS: default setting is two-tail test u<> 5 HOW TO DO? Compare mean 1sample t-test enter test value = 5 test value = 5 == U=5

Independent samples t-test

Ho: UM = UF H1a: UM <> UF HOW TO DO? Alayze compare means independent sample t-test Test variable : likihood of buying; grouping variable: gender define group: group 1:1 ; group 2=2

Step 1: check p-value of Levenes test. If p-value of Levenes test >= 0.05, use the Equal variances assumed row, otherwise, use the p-value of Equal cariances not assumed row.
Independent Samples Test Levene's Test for Equality of Variances F likelihood of buying Equal variances assumed Equal variances not assumed Lower .370 Sig. Upper .549 t Lower -.021 -.021 df Upper 23 22.281

t-test for Equality Sig. (2-tailed) Lower .984 .983

Mean Diffe Upper

Since Levenes test p-value = 0.549 > 0.05, therefore we use the t-test p-value of Equal variances assumed row to test our hypotheses Since p-value = 0.984 > 0.05, we do not reject H1a. There is no gender difference in buying intention. There is no need to formulate different market mix strategies for the two gender groups. Gender is not a good variable to segment the market.

Paired samples t-test

Repeated measurements taken from the same subject H2: d = 0 (d = difference between two measurements, which are aware_b, aware_a) H2: d<> 0 Difference : making a change HOW TO DO? Analyze compare means paired sample t-test highlight 2 variables Click button
Paired Samples Test Paired Differences Mean Lower Pair 1 Awareness Level before Viewing the Ad - Awareness Level after Viewing the Ad -6.50000 Std. Deviation Upper 15.14476 Std. Error Mean Lower 95% Confidence Interval of the Difference Lower -16.12252 Upper 3.12252 t Mean df Std. Deviation Lower -1.487 Upper 11

Sig. (2-tail

Std. Err

Upper 4.37191

Since p-value = 0.165 > 0.05, do not reject H2. There is no significant improvement in awareness after viewing the new ad. New ad is useless in improving customers level of awareness.

Crosstab
X Nominal Nominal >=ordinal >=ordinal X = gender Y = shopping duty H3: gender and shopping duty are independent H3a: gender and shopping duty are dependent Y Nominal >=Ordinal Nominal >=ordinal Method Crosstab (ch-sq test) ANOVA Discriminant Analysis Regression

HOW TO DO?
Analyze description crosstabs Row: gender; Column: shopping sales stat: tick chi=square, contingency coefficient, Phi and Cramers V; cell: expected and ordinance Phi & Cramers V= r2 P(M) = 14/25; P(shared duty) = 8/25 If gender and duty are independent, P(M & share duty) = 14/25 * 8/25 In n = 25, the expected # of people who one M & share duty = 14/25 * 8/25 * 25=4.5
gender * shopping duty Crosstabulation shopping duty Share duty 4 3.5 4 4.5 8 8.0 Total Yes 3 4.4 7 5.6 10 10.0 11 11.0 14 14.0 25 25.0

Yes gender female male Total Count Expected Count Count Expected Count Count Expected Count 4 3.1 3 3.9 7 7.0

No

3.1 = 7*11/25

Chi-Square Tests Asymp. Sig. Value df (2-sided) Pearson Chi-Square 1.403(a) 2 .496 Likelihood Ratio 1.428 2 .490 Linear-by-Linear Association 1.260 1 .262 N of Valid Cases 25 a 5 cells (83.3%) have expected count less than 5. The minimum expected count is 3.08. Use the Pearson chi-sq p-value if the n >= 30 Use the Likelihood tation p-value for small smaple; n<=30 Since p-value = 0.490 > 0.05, do not reject H3. There is no gender difference in shopping duty requirement.

If a

5 cells (83.3%) > 20% unreliable

Since the sample size is small Solution: compare Yes and share duty group result: 2 by 2 table HOW to DO??? Transform, recode into different variables shopping click old and new rangeRANGE 1 and 2 ; value:1 value:3, value: 2 gender * recoded_duty Crosstabulation recoded_duty 1.00 2.00 8 3 6.6 4.4 7 7 8.4 5.6 15 10 15.0 10.0 Chi-Square Tests Asymp. Sig. (2-sided) .250 .459 .245 Exact Sig. (2-sided) Exact Sig. (1-sided) Total 1.00 11 11.0 14 14.0 25 25.0

gender

female male

Total

Count Expected Count Count Expected Count Count Expected Count

Pearson Chi-Square 1 Continuity Correction(a) 1 Likelihood Ratio 1 Fisher's Exact Test Linear-by-Linear Association 1.273 1 .259 N of Valid Cases 25 a Computed only for a 2x2 table b 1 cells (25.0%) have expected count less than 5. The minimum expected count is 4.40.

Value 1.326(b) .548 1.352

df

.414

.231

** Fishers Exact Test can be used even cell > 20% & you cannot reduce your sample size further. If symmetric measures < 0.3, the correlation relationship between X(gender) & Y(duty) is weak.

Crossbab.spss H4: gender (X) and made purchase (Y) are independent H4a: gender and made purchase are dependent
Chi-Square Tests Asymp. Sig. (2-sided) .000 .000 .000 Exact Sig. (2-sided) Exact Sig. (1-sided)

Pearson Chi-Square 1 Continuity Correction(a) 1 Likelihood Ratio 1 Fisher's Exact Test Linear-by-Linear Association 18.608 1 .000 N of Valid Cases 582 a Computed only for a 2x2 table b 0 cells (.0%) have expected count less than 5. The minimum expected count is 70.38.

Value 18.640(b) 17.859 18.368

df

.000

.000

N = 582, use Peasrson chi-sq p-value = 0.000 <0.05, reject N4. Gender and made purchase are dependent. There is a gender difference in made purchase. ** Check pattern by examining the standardized residuals

>2?

gender * made purchase Crosstabulation made purchase no yes 94 115 70.4 138.6 2.8 -2.0 102 271 125.6 247.4 -2.1 1.5 196 386 196.0 386.0 Total no 209 209.0 373 373.0 582 582.0

gender

male

female

Total

Count Expected Count Std. Residual Count Expected Count Std. Residual Count Expected Count

Male customers are nonbuyers (2.8); female customers tend to be buyers(1.5). we should target female customers more.

Regression analysis
X: age; Y = prob *nominal variables can be used if it comprises 2 groups only. Transform compute variable age: = 2012-birth year

Anlayze regression liner

Coefficients(a) Standardized Coefficients Beta -.287 -.103

Unstandardized Coefficients B Std. Error (Constant) 10.408 4.487 age -.090 .072 gender -.617 1.391 a Dependent Variable: likelihood of buying Model 1

t B 2.320 -1.241 -.444

Sig. Std. Error .031 .229 .662

prob^ = 10.408 0.9*age -0.617*gender ^=predicted value Whats the predicted prob of a male customer aged 50? prob^ = 10.408 0.9*50-0.617*1= The absolute Beta value represents the importance of the independent variable in affecting the dependent variable. Age is more important than gender. Prob: ZRESID Y; ZPRED X

Sc a t t e r p l o t

Re g r e s s i o n St a n d a r d i z e d Re s i d u a l

De p e n d e n t Va r i a b l e : l i k e l i h o o d o f b u y i n g
2

-1

-2 -2 -1 0 1 2

Re g r e s s i o n St a n d a r d i z e d Re s i d u a l

H5: beta of age = 5 H5a: beta of age <> 5 Since p-value =0.229 > 0.05, we do not reject H5. There is no age difference in buying intention. H6: beta of gender = 5 H6a: beta of gender <> 5 Since p-value =0.662 > 0.05, we do not reject H6. There is no gender difference in buying intention. WITHOUT gender
Coefficients(a) Standardized Coefficients Beta -.250

Unstandardized Coefficients (Constant) 8.911 age -.078 a Dependent Variable: likelihood of buying Model 1 B Std. Error 2.898 .066

t B 3.075 -1.182

Sig. Std. Error .006 .251

** Residuals : SSE (error of regression) Benchmark, standard model Yi = -Y SST=197.652 SSE: Z (Yi-Yi^) = 185.330 SSi- SSE = SSR = 12.32
ANOVA(b) Sum of Squares Regression 12.322 Residual 185.330 Total 197.652 a Predictors: (Constant), age b Dependent Variable: likelihood of buying Model 1 df 1 21 22 Mean Square 12.322 8.825 F 1.396 Sig. .251(a)

H7: All beta coefficients of the equation = 0 H7a: All beta coefficients of the equation <>0
Coefficients(a) Standardized Coefficients Beta -.287 -.103

Unstandardized Coefficients B Std. Error (Constant) 10.408 4.487 age -.090 .072 gender -.617 1.391 a Dependent Variable: likelihood of buying Model 1

t B 2.320 -1.241 -.444

Sig. Std. Error .031 .229 .662

ANOVA(b) Sum of Squares Regression 14.127 Residual 183.525 Total 197.652 a Predictors: (Constant), gender, age b Dependent Variable: likelihood of buying Model 1 df 2 20 22 Mean Square 7.063 9.176 F .770 Sig. .476(a)

From ANOVA table, p-value =0.476 > 0.05, do not reject H7. The regession model/ equation is useless.