(2020/2021) SPSS Psychometrie (uitwerkingen met stappenplan)

WEEK 1 METEN, SCHALEN EN NORMEN

2. Calculate both scale scores for the LOK by finding the mean of the items concerned
(see Annex A1, A3). Remember that your choice of a value for this ‘n’ has
consequences for the quality of the scale scores! What does a high value on the mean
LOK scale score indicates for a child?
1. Annex 1 → Emotional neglect: items 1, 2, 3, 4, 5, 6, 7, 9, 10, 11
2. Transform → Compute variable
- Target Variable: M​ EAN.LOKf
- Numeric expression:​ Mean.5(lokfa1 to lokfa7,lokfa9 to lokfa11)
- Target Variable:​ MEAN.LOKm
- Numeric expression: ​Mean.5(lokmo1 to lokmo7,lokmo9 to lokmo11)

* MEAN.N (drempel minimum aantal items dat ingevuld is)




3. Create a report about the research group: Provide relevant statistics separately for
each of the scales across the total research group.
1. Analyze → Descriptive Statistics → Descriptives
2. ​Variable(s):​ MEAN.LOKf, MEAN.LOKm
3. Opt​: request mean, standard deviation, minimum, and maximum
4. Execute




1. Graphs → Legacy dialogs → Histogram
2. ​Variable(s):​ MEAN.LOKf
3. Execute




1

,1. Graphs → Legacy dialogs → Histogram
2. ​Variable(s):​ MEAN.LOKm
3. Execute




4. Creating norms of the scales (see Furr (2018, pp. 56 – 68)): Calculate normed
scores for the scales of the LOK:

a. Calculate z-scores.
1. Analyze → Descriptive Statistics → Descriptives
2. ​Variable(s):​ MEAN.LOKf, MEAN.LOKm
3. Save standardized variables
4. Execute




b. Calculate T-scores
1. Transform → Compute variable
​ MEAN.LOKf
- Target Variable: T
- Numeric expression:​ 50 + 10 * ZMEAN.LOKf

2. Transform → Compute variable
​ MEAN.LOKm
- Target Variable: T
- Numeric expression:​ 50 + 10 * ZMEAN.LOKm




2

,c. Get SPSS to determine percentile scores
1. Transform → Rank Cases
2. ​Variable(s):​ MEAN.LOKm, MEAN.LOKf
3. ​Rank Types:
- Rank ​OFF
- Fractional rank as %’’ ​ON
4. ​Ties​: High




*change P.MEAN.LO to PMEAN.LOKf
*change PER001 to PMEAN.LOKm

5. Create norm table: For each scale of the LOK, create a norm table (table of raw
scores [without repeats] + the three different normed scores).
For​ each ​scale of the LOK, create a norm table.

Normtabel Lokf
1. Analyze → Reports → Case Summaries
2. ​Variable(s):​ norm scores (ZMEAN.LOKf, TMEAN.LOKf, PMEAN.LOKf)
3. ​Grouping variables:​ scare score (MEAN.LOKf)
4. Display cases ​OFF
5. ​Statistics​ → Cell statistics : mean




3

, Normtabel Lokm
1. Analyze → Reports → Case Summaries
2. ​Variable(s):​ norm scores (ZMEAN.LOKm, TMEAN.LOKm, PMEAN.LOKm)
3. ​Grouping variables:​ scare score (MEAN.LOKm)
4. Display cases ​OFF
5. ​Statistics​ → Cell statistics : mean




6. Check whether the norm tables are ‘correct’ by
looking at minima and maxima of the various
normed score distributions.
• Gemiddelde Z-score is 0 : ​CORRECT
• Gemiddelde T-score is 50 : ​CORRECT
• Percentielscores tellen op tot 100 : ​CORRECT




4