ch7

Question 1

What is archival data

Answer 1

• Data gathered from existing sources
• Collected for another purpose than that of the current study
• Secondary data (i.s.o. primary)

Beware: existing survey data = secondary data

Question 2

Internal archival data

Answer 2

Company records and archives

Question 3

External archival data

Answer 3

• Publicly available data

* Commercially available data

Question 4

What are the strengths of archival research:

Answer 4

1. Tapping into industry wisdom
2. Examining effects across time
3. Examining effects across countries
4. Examining socially sensitive phenomena

Question 5

Piecing together archival data

Answer 5

•Unit of analysis
= level of data analysis/data aggregation
= level at which DV is measured

•Make sure your unit of analysis correspondents
IV measured at same level or higher level than DV

Question 6

Unit of Analysis levels

Answer 6

• Country
• Industry
• Firm
• Brand
• Consumer

Question 7

Unit of Analysis levels (longitudinal data, time-variant)

Answer 7

• Year
• Quarter
• Month
• Week
• Day

Question 8

Sources of measurement unreliability in archival research

Answer 8

• Missing data
• Inaccurately recorded data
• Fake data

Question 9

Combining multiple archival indicators into a single measure

Answer 9

1) Standardise each indicator

2) Average the standardised indicators

Question 10

Solutions to missing data in cross-sectional data sets (meaning without time dimension)

Answer 10

• Listwise deletion; delete entire row if 1 variable is missing *check user manual
• Mean substitution: replace missing value for observation I and variable j with average value on variable j for all other observations

Question 11

Solution to missing data in longitudinal data sets

Answer 11

Interpolation

Question 12

Solution to inaccurately recorded data (longitudinal data set)

Answer 12

•Inaccuracies that turn up as extreme data points

SOLUTION: Remove observation: run analyses with and without observation

•Inaccuracies that are not extreme

SOLUTION: Trim/truncate (in large data sets). Remove a fraction of observations, e.g. 1% most extreme observations

*Plot data to notice extreme data points

Question 13

Fake observations

Answer 13

• Be critical!
• Who collected the data?
• When? Where?
• For what purpose?

Question 14

Measurement validity

Answer 14

Does a measure represent the construct it is supposed to measure?

Considerable conceptual overlap: valid measure
Little conceptual overlap: questionable proxy

An archival measure may only be a “proxy” of the underlying construct (approximation of construct)

Question 15

How to validate than an archival measure is a good measure rather than a bad proxy?

Answer 15

• Provide precedence (referring to high quality studies)
• Provide sound logic to support that considerable conceptual overlap exists between construct and measure
• Provide evidence of a substantial correlation between your proxy and a valid survey measure for a small subsample of your data
• provide evidence of substantial correlations (r>.3) with related constructs (“nomological validity”)

Question 16

Generalisability (external validity)

Answer 16

Is the sample representative of the population? Always read the documents explaining the methodology behind archival databases

Golden rule of archival research: IV at the same or higher level than DV
DV at firm level + IV at industry level: possible
DV at industry level + IV at firm level: not possible