Concepts to be included Causality Time order see

Concepts to be included § § § Causality Time order (see: reverse causation) Association (between variables) Spurious relationship Cross-sectional research

CROSS-SECTIONAL RESEARCH HENK VAN DER KOLK

AIM Introducing cross-sectional research Some threats to causal inference in cross-sectional research

CROSS-SECTIONAL RESEARCH (DEF. ) A research design, in which. . . all variables of a set of units are measured at the same time and none of the variables is manipulated differently for a sub-set of units.

EXAMPLE Does the amount of e-marketing increase sales? Amount of e-marketing + Sales Companies (? ) Months in one company (? )

EXAMPLE Collecting data about a set of online shops … Asking both e-marketing budgets and sales.

THREE ASPECTS OF CAUSALITY Sales Association; X and Y are correlated Amount of e-marketing spending

THREE ASPECTS OF CAUSALITY Association; X and Y are correlated Time order; X precedes Y in time Non spuriousness; no other variable (Z) produces the correlation.

ASSESSING CAUSAL RESEARCH I. Measurement validity (and reliability) II. External validity III. Internal validity IV. Statistical conclusion validity Time order Non-spuriousness Correlation

MAIN THREATS TO INTERNAL VALIDITY IN CROSS-SECTIONAL RESEARCH Reverse causation cannot be ruled out Third variables may affect the relationship (relationship may be spurious)

REVERSE CAUSATION Example: Sales increases the budget for e-marketing Amount of e-marketing spending Since data are collected at one moment in time, reverse causation cannot be ruled out Sales

MEASUREMENT BIAS MAY PRODUCE REVERSE CAUSATION Sometimes, the problem of reverse causation can be reduced in cross-sectional research. Example: ask for the budget of last year, and this years sales.

MEASUREMENT BIAS PRODUCES REVERSE CAUSATION Measuring both variables at the same time sometimes produces reverse causation. Example: Does a ‘happy childhood’ make you a more ‘happy adult’?

THIRD VARIABLES (SPURIOUSNESS) Third variables may affect the relationship (relationship may be spurious) Example: Maybe the presence of ‘young and dynamic managers’ (Z) affect both sales (Y) and the e-marketing budget (X)?

CONFOUNDING Number of young and dynamic managers E-marketing budget Sales 15

THIRD VARIABLES (SPURIOUSNESS) ‘Confounding’ is a poblem in cross-sectional research. Taking into account the possible effect of third variables, may reduce the problems of spuriousness. ‘Controlling’ for third variables.

EVALUATING CROSS-SECTIONAL RESEARCH Weak in internal validity (reverse causation / third variables) Potentially strong in external validity (sampling) The effect of many independent variables cannot be studied in other types of research designs

AIM Introducing cross-sectional research Some threats to causal inference in cross-sectional research

IMAGES USED: § Slide 4: https: //pixabay. com/en/e-commerce-shopping-basket-shopping-402822/ § Slide 5: https: //pixabay. com/en/iphone-visa-business-buying-card-624709/ § Slide 9: https: //pixabay. com/en/personal-group-silhouettes-man-365964/