1

The Midlife Development Inventory (MIDI)

Personality Scales: Scale Construction and Scoring

Margie )* Lachman 0 Su3anne L* 4eaver

Technical 8eport < =uly 1??@

Arandeis Dniversity

Esychology DepartmentG MS H62

4althamG MA H2454N?11H

8esearch supported by The =ohn D* and Patherine T* MacArthur Qoundation

8esearch Network on Successful Midlife Development VMIDMAPX*
 2

The MIDI includes 6 personality trait scales: NeuroticismG )xtraversionG [penness to

)xperienceG PonscientiousnessG Agreeableness VPommunionXG and Agency* These traits were

measured using selfNratings of 3H ad]ectives* Most of the ad]ectives were selected from existing

trait lists and inventories Vsee AemG 1?^1_ GoldbergG 1??2_ =ohnG 1??H_ Trapnell 0 4igginsG

1??HX_ a few items were generated by Margie Lachman and Alice 8ossi* Qor each personality

dimensionG a list of all ad]ectives appearing in the literature was compiled V=ohnG 1??HX* The

ad]ectives which appeared most consistently as markers and had the highest item to total

correlations or factor loadings were identified* InitiallyG four ad]ectives were selected to mark

each trait dimension* Given the time constraints for MIDIG the goal was to create the shortest

possible set of items to measure personality reliably in less than five 5 minutes by telephone or

mail survey*

A pilot study was conducted in 1??4 with a probability sample of 1HHH men and women

between the ages of 3H and @H* Items with the highest item to total correlations and factor

loadings were selected from MIDI* Qorward regressions were also run to determine the smallest

number of items needed to account for over ?Ha of the total scale variance* Many of the

negatively worded items VunemotionalG unreliableG unsophisticatedG unsympatheticG shyG

unsociableX were dropped due to very low variance* New items were added to increase

reliabilities on some scales*

Information about the factor loadingsG coefficient alphaG item to total correlations and

scale means from the MIDDS sampleG and the scoring algorithm is contained in this document*
 3

REFERENCES

AemG S* L* V1?^1X* !"# %"&'()*" +,-",.)/0 12,32*4 Ealo AltoG PA: Ponsulting Esychologists
Eress*

GoldbergG L* 8* V1??2X* The development of markers for the AigNQive factor structure*
56078)*)9:72* ;66"66#",.< =G 26N42*

=ohnG [* E* V1??HX* The bAig Qivec factor taxonomy: Dimensions of personality in the natural
language and in duestionnaires* In L*A* Eervin V)d*XG >2,?@))A )B C"/6),2*:.0 .8")/0
2,? /"6"2/78< Vpp* 66N1HHX* New eork: Guilford*

TrapnellG E* D*G 0 4igginsG =* S* V1??HX* )xtension of the Interpersonal Ad]ective Scales to

include the Aig Qive dimensions of personality* D)3/,2* )B 5"/6),2*:.0 2,? %)7:2*
56078)*)90< EF< @^1N@?H
 4

Research Using MIDI Personality Items

Mroc3ekG D*G 0 folar3G P* V1??^X* The effect of age on positive and negative affect: A
developmental perspective on happiness* D)3/,2* )B 5"/6),2*:.0 2,? %)7:2* 56078)*)90<
@5G 1333N134?*

ElautG g*G MarkusG h*G 0 LachmanG M* V2HH2X* Elace matters: Ponsensual features and regional
variation in American wellNbeing and self* D)3/,2* )B 5"/6),2*:.0 2,? %)7:2*
56078)*)90< ^3G 16HN1^4*

ErendaG f* 0 LachmanG M* V2HH1X* Elanning for the future: A life management strategy for
increasing control and life satisfaction in adulthood* 56078)*)90 2,? ;9:,9< 16G 2H6N216*

StaudingerG D*G QleesonG 4*G 0 AaltesG E* V1???X* Eredictors of sub]ective physical health and
global wellNbeing: Similarities and differences between the Dnited States and Germany*
D)3/,2* )B 5"/6),2*:.0 2,? %)7:2* 56078)*)90< @6G 3H5N31?*
 5

MIDI

The Midlife Development Inventory

Part I Duestionnaire

The =ohn D* and Patherine T* MacArthur Qoundation

8esearch Network [n Successful Midlife Development
6
 @

Scoring fey for MIDI Eersonality Trait Scales

Eart I iuestionnaireG Section Q4G p*34

1* Naming of items: An bsc prefix was added to all items* V)xample: b[utgoing is item f4a
and is named sf4a*X

2* Items are reverse scored* An brc suffix is added to the reversed item* All items are
reversed )jP)ET sf4s and sf4x*

3* Scales are computed by finding the mean of the items for cases that have valid values for
at least half of the items on the particular scale* System missing values should be
assigned for cases with fewer than half the items for the scale* Qor scoring the
personality scales the number in the computer code specifies the minimum number of the
items reduired* Qor example M)AN *3 signifies for the agency scale at least 3 items
should be completed to count the case as nonNmissing*

4* Qor all scalesG higher values indicate greater endorsement of the construct*

8ecode sf4a sf4b sf4c sf4d sf4e sf4f sf4g sf4h sf4i sf4] sf4k sf4l
Sf4m sf4n sf4o sf4p sf4d sf4r sf4t sf4u sf4v sf4w sf4y sf43
Sf4aa sf4bb sf4cc sf4dd
V1k4X V2k3X V3k2X V4k1X into
sf4ar sf4br sf4dr sf4er sf4fr sf4gr sf4hr sf4ir sf4]r sf4kr sf4]r
sf4mr sf4nr sf4or sf4pr sf4dr sf4rr sf4tr sf4ur sf4vr sf4wr sf4yr
sf43r sf4aar sf4bbr sf4ccr sf4ddr

Pompute agency k mean*3Vsf4erG sf4]rG sf4HrG sf4trG sf4ddrX

Pompute agree k mean*3Vsf4brG sf4grG sf4lr sf4rrG sf43rX
Pompute open k mean*4Vsf4nrG sf4drG sf4urG sf4vrG sf4yrG sf4bbrG sf4ccrX
Pompute nurot k mean*2Vsf4crG sf4hrG sf4mrG sf4sX
Pompute extrav k mean*3Vsf4arG sf4frG sf4krG sf4wrG sf4aarX
Pompute consc k mean*2 Vsf4drG sf4irG sf4prG sf4xX

gariable labels
agency lsf4 agencym
agree lsf4 agreeablenessm
neurot lsf4 neuroticismm
open lsf4 openness to experiencem
consc lsf4 conscientiousnessm
extrav lsf4 extraversionm
^
?