You are on page 1of 10

Use of a data model (single version of the truth)

Yes 277 201 72.6%


No 277 76 27.4%
Determinants of the use of data analytics in project situations

Deal complexity 277 201 72.6%


Deal scope 277 202 72.9%
Compatibility with client’s IT capabilities 277 70 25.3%
Time restrictions 277 119 43.0%
Client’s demand for interim results 277 38 13.7%
Budget restrictions 277 116 41.9%
Availability of skilled resources 277 91 32.9%
Data availability 277 244 88.1%
Data variety 277 91 32.9%
Data veracity 277 63 22.7%
Profitability analyses benefitting from data analytics

Cohort analysis 277 96 34.7%


Customer churn analysis 277 112 40.4%
Customer lifetime value (CLTV) analysis 277 54 19.5%
Identification of one-offs 277 59 21.3%
Price-volume analysis 277 183 66.1%
Raw material pass-through analysis 277 21 7.6%
Reconciliations 277 64 23.1%
Sum of the parts P&L 277 77 27.8%
Transaction effect analysis 277 29 10.5%
Translation effect/constant currency analysis 277 46 16.6%

Notes:
1) The frequency displays the sum of top 1, top 2, and top 3 listings of the corresponding items.

Source: Own illustration based on survey results


Table A.4: Use of data analytics – Summary statistics of continuous variables
Variable n Mean Median Std. deviation Min. Max.
Impact and suitability

Impact 330 4.49 5 0.65 2 5


Suitability 320 4.16 4 0.79 1 5
Determinants of data availability and granularity

Initiator 280 4.50 5 0.74 2 5


Negotiation situation 265 3.35 3 1.05 1 5
Target company size 285 3.89 4 1.02 1 5
Selling party 266 4.04 4 0.95 1 5
Public listing 252 3.05 3 1.22 1 5
Financial sponsor ownership 265 3.60 4 1.03 1 5
Sales fragmentation 254 3.05 3 1.10 1 5
Data culture 269 4.29 4 0.82 1 5
Usage of different data types

Target-internal financial data 287 4.90 5 0.39 2 5


Target-internal non-financial data 285 3.94 4 0.84 1 5
Target-external financial data 286 3.05 3 1.04 1 5
Target-external non-financial data 286 2.87 3 1.02 1 5
Time shifts in data preparation and analysis

Time shifts 270 4.13 4 1.03 1 5


Emerging trade-off 272 4.21 5 1.03 1 5
Use of data analytics across review areas

Profitability analysis (incl. quality of earn.) 268 3.94 4 0.88 1 5


Balance sheet analysis 265 3.37 4 1.14 1 5
Cash flow analysis 265 3.13 3 1.12 1 5
Business plan validation 255 3.11 3 1.10 1 5
Future technological developments

Full automation of key analyses 264 3.66 4 1.22 1 5


Global benchmarking database 264 3.89 4 1.06 1 5
Machine learning-based classification 265 3.65 4 1.08 1 5
Predictive analytics-based business plan 259 3.67 4 1.04 1 5
Interactive dashboards 262 4.55 5 0.75 1 5
Impact of data analytics on audit firms

Maximization of efficiency gains 257 4.25 4 0.74 1 5


Shift towards value and insight-orientation 258 4.31 4 0.69 2 5
Use of cross-functional teams 257 4.08 4 0.83 1 5
Alternative pricing approaches 244 3.49 4 1.15 1 5
Impact of data analytics on FDD in the M&A process

Increasing links to other due diligence forms 263 4.16 4 0.83 2 5


Increasing integration into M&A process 261 4.13 4 0.90 1 5
Source: Own illustration based on survey results
Appendix 6: Measurement model – Parameter estimates
Table A.5: Measurement model – Unstandardized parameter estimates
Variable Expected sign Unstd. coefficient Std. error z-value p-value
Actual use (USE)

USE11
USE + 1.00003

Constant + 3.9944 0.1050 38.04 0.000****


Behavioral intention (BI)

BI1
BI
+ 1.00003
Constant + 5.9133 0.0753 78.56 0.000****
BI2
BI
+ 1.1295 0.0450 25.10 0.000****
Constant + 5.7905 0.0816 70.93 0.000****
BI3
BI
+ 1.1300 0.0502 22.52 0.000****
Constant + 5.7501 0.0855 67.28 0.000****
Performance expectancy (PE)

PE1
PE
+ 1.00003

Constant + 5.2849 0.0921 57.37 0.000****


PE2
PE
+ 0.7230 0.0708 10.21 0.000****
Constant + 5.7290 0.0700 81.81 0.000****
Effort expectancy (EE)

EE1
EE
+ 1.00003
Constant + 4.6430 0.9693 47.90 0.000****
EE2
EE
+ 0.9799 0.0431 22.72 0.000****
Constant + 4.4579 0.0937 47.60 0.000****
EE3
EE
+ 0.9439 0.0404 23.38 0.000****
Constant + 4.6081 0.8945 51.52 0.000****
Social influence (SI)

SI1
SI
+ 1.00003
Constant + 5.3339 0.0878 60.74 0.000****
SI2
SI
+ 0.7299 0.0733 9.96 0.000****
Constant + 5.2932 0.8964 59.05 0.000****
SI3
SI
+ 0.8249 0.0993 8.30 0.000****
Constant + 4.4463 0.1111 40.02 0.000****
Facilitating conditions (FC)

FC1
FC
+ 1.00003
Constant + 5.0275 0.0993 50.63 0.000****
FC2
FC
+ 1.3943 0.1019 13.69 0.000****
Constant + 4.7354 0.1011 46.85 0.000****
FC32
FC + 0.7972 0.0864 9.23 0.000****
Constant + 5.2107 0.0961 54.22 0.000****

Notes:
Error variances of the indicator variables, variances of the latent variables, and covariances between the latent variables are not
displayed.
1) The item was requested on a scale from 0 to 100 and has been linearly transformed to a scale from 1 to 7.
2) The item refers to a negative-wording question and has therefore been reverse-coded.
3) The value is pre-determined to 1 since the item serves as the reference indicator. Consequently, no standard error and z-
statistics are displayed.

Variables definition:
, , ,
* ** *** ****
indicate p<0.1, 0.05, 0.01, and 0.001 respectively

Source: Own illustration based on survey results


Table A.6: Measurement model – Standardized parameter estimates
Variable Expected sign Std. coefficient Std. error z-value p-value
Actual use (USE)

USE11
USE + 1.00003 0.0000 4.2e16 0.000****
Constant + 2.3151 0.1166 19.85 0.000****
Behavioral intention (BI)

BI1
BI
+ 0.9050 0.0141 64.30 0.000****
Constant + 4.8116 0.2186 22.01 0.000****
BI2
BI
+ 0.9418 0.0107 88.10 0.000****
Constant + 4.3410 0.1983 21.90 0.000****
BI3
BI
+ 0.9002 0.0140 64.09 0.000****
Constant + 4.1190 0.1904 21.63 0.000****
Performance expectancy (PE)

PE1
PE
+ 0.7929 0.0385 20.62 0.000****
Constant + 3.5270 0.1679 21.01 0.000****
PE2
PE
+ 0.7527 0.0394 19.11 0.000****
Constant + 5.0198 0.2280 22.01 0.000****
Effort expectancy (EE)

EE1
EE
+ 0.9024 0.0145 62.02 0.000****
Constant + 2.9578 0.1449 20.41 0.000****
EE2
EE
+ 0.9153 0.0133 69.07 0.000****
Constant + 2.9394 0.1445 20.35 0.000****
EE3
EE
+ 0.9242 0.0124 74.25 0.000****
Constant + 3.1852 0.1552 20.53 0.000****
Social influence (SI)

SI1
SI
+ 0.8504 0.0327 25.98 0.000****
Constant + 3.7839 0.1815 20.85 0.000****
SI2
SI
+ 0.6073 0.0459 13.23 0.000****
Constant + 3.6743 0.1741 21.11 0.000****
SI3
SI
+ 0.5542 0.0498 11.12 0.000****
Constant + 2.4920 0.1284 19.41 0.000****
Facilitating conditions (FC)
FC1
FC
+ 0.7099 0.0345 20.60 0.000****
Constant + 3.0969 0.1481 20.92 0.000****
FC2
FC
+ 0.9710 0.0166 58.56 0.0000****
Constant + 2.8616 0.1380 20.73 0.0000****
FC32
FC + 0.6058 0.0442 13.69 0.0000****
Constant + 3.4359 0.1727 19.90 0.0000****

Notes:
Error variances of the indicator variables, variances of the latent variables, and covariances between the latent variables are not
displayed. Variances of the latent variables are fixed to zero for the calculation of standardized coefficients.
1) The item was requested on a scale from 0 to 100 and has been linearly transformed to a scale from 1 to 7.
2) The item refers to a negative-wording question and has therefore been reverse-coded.
3) The value is pre-determined to 1 since the actual use (USE) construct represents a single-indicator latent variable for which
the error variance is artificially fixed to zero for model identification purposes (i.e., it technically equals a manifest variable).

Variables definition:
, , ,
* ** *** ****
indicate p<0.1, 0.05, 0.01, and 0.001 respectively

Source: Own illustration based on survey results


Appendix 7: Structural model – Parameter estimates
Table A.7: Structural model – Standardized parameter estimates
Variable Expected sign Std. coefficient Std. error z-value p-value
Effects on behavioral intention (BI)

H1: PE → BI + 0.2813 0.0914 3.08 0.002***


H2: EE → BI + 0.0294 0.0715 0.41 0.681
H3: SI → BI + 0.6387 0.0892 7.16 0.000****
Effects on actual use (USE)

H4: FC → USE + 0.2161 0.0725 2.98 0.003***


H5: BI → USE + 0.4659 0.0695 6.71 0.000****

Notes:
Variances of the latent variables and covariances between the latent variables are not displayed. Variances of the latent variables
are fixed to zero for the calculation of standardized coefficients.

Variables definition:
, , ,
* ** *** ****
indicate p<0.1, 0.05, 0.01, and 0.001 respectively

Source: Own illustration based on survey results

Table A.8: Structural model – Standardized parameter estimates (GSANL subsample)


Variable Expected sign Std. coefficient Std. error z-value p-value
Effects on behavioral intention (BI)

H1: PE → BI + 0.2223 0.1041 2.13 0.033**


H2: EE → BI + 0.0618 0.0874 0.71 0.480
H3: SI → BI + 0.6813 0.1119 6.09 0.000****
Effects on actual use (USE)

H4: FC → USE + 0.2093 0.0807 2.59 0.010**


H5: BI → USE + 0.4348 0.0777 5.60 0.000****

Notes:
Variances of the latent variables and covariances between the latent variables are not displayed. Variances of the latent variables
are fixed to zero for the calculation of standardized coefficients.

Variables definition:
, , ,
* ** *** ****
indicate p<0.1, 0.05, 0.01, and 0.001 respectively

Source: Own illustration based on survey results


Appendix 8: Multi-group analysis of interaction effects
Table A.9: Analysis of the moderating effect of gender
(Pot.) moderated Likelihood
Total sample Group samples
effect ratio test
n=262 Females, n=61 Males, n=201
Mean diff. [0-1] χ2 (df)
Unstd. coefficient Unstd. coefficient[0] Unstd. coefficient[1]
H1a: PE → BI 0.3030*** 0.2533* 0.3157*** -0.4679** 0.26 (1)
H2a: EE → BI 0.0245 -0.0593 0.0584 -0.1958 1.93 (1)
SI → BI 0.5668**** 0.4256**** 0.6116**** -0.4671** 2.80 (1)*
FC → USE 0.3372*** 0.3301* 0.3393*** -0.1534 0.00 (1)

Notes:
For each potentially moderated effect, a model with measurement equivalence constraints and a parallel slopes assumption for
all latent variables was tested against the almost identical model for which the parallel slopes assumption was released for the
potentially moderated effect. Consequently, the likelihood ratio test that compares whether the models are significantly different
always has one degree of freedom.

Variables definition:
* ** *** ****
, , , indicate p<0.1, 0.05, 0.01, and 0.001 respectively

Source: Own illustration based on survey results

Table A.10: Analysis of the moderating effect of age


(Pot.) moderated Likelihood
Total sample Group samples
effect ratio test
n=248 Young, n=114 Old, n=134
Mean diff. [0-1] χ2 (df)
Unstd. coefficient Unstd. coefficient[0] Unstd. coefficient[1]
H1b: PE → BI 0.2750** 0.2389* 0.3100** -0.2119 0.61 (1)
H2b: EE → BI -0.0226 -0.0762 0.0116 -0.6521**** 1.41 (1)
SI → BI 0.6137**** 0.4850*** 0.6369**** -0.3000* 2.18 (1)
FC → USE 0.2935*** 0.4090*** 0.1720 -0.3284** 2.56 (1)

Notes:
For each potentially moderated effect, a model with measurement equivalence constraints and a parallel slopes assumption for
all latent variables was tested against the almost identical model for which the parallel slopes assumption was released for the
potentially moderated effect. Consequently, the likelihood ratio test that compares whether the models are significantly different
always has one degree of freedom.

Variables definition:
, , ,
* ** *** ****
indicate p<0.1, 0.05, 0.01, and 0.001 respectively

Source: Own illustration based on survey results


Table A.11: Analysis of the moderating effect of hierarchy level (1/3)
(Pot.) moderated Likelihood
Total sample Group samples
effect ratio test
n=265 Junior, n=177 Senior, n=88
Mean diff. [0-1] χ2 (df)
Unstd. coefficient Unstd. coefficient[0] Unstd. coefficient[1]
H1c: PE → BI 0.2495*** 0.2313** 0.2690*** -0.3118 0.17 (1)
H2c: EE → BI 0.0144 -0.0138 0.0689 -0.7415**** 1.27 (1)
SI → BI 0.6037**** 0.5735**** 0.6668**** -0.1849 0.96 (1)
FC → USE 0.3386*** 0.3607*** 0.3001* -0.4285*** 0.14 (1)

Notes:
For each potentially moderated effect, a model with measurement equivalence constraints and a parallel slopes assumption for
all latent variables was tested against the almost identical model for which the parallel slopes assumption was released for the
potentially moderated effect. Consequently, the likelihood ratio test that compares whether the models are significantly different
always has one degree of freedom.

Variables definition:
, , ,
* ** *** ****
indicate p<0.1, 0.05, 0.01, and 0.001 respectively

Source: Own illustration based on survey results

Table A.12: Analysis of the moderating effect of hierarchy level (2/3)


(Pot.) moderated Likelihood
Total sample Group samples
effect ratio test
n=248 Junior, n=177 Senior, n=71
Mean diff. [0-1] χ2 (df)
Unstd. coefficient Unstd. coefficient[0] Unstd. coefficient[1]
H1c: PE → BI 0.2221** 0.2056* 0.2416** -0.3558* 0.13 (1)
H2c: EE → BI 0.0121 -0.0085 0.1073 -0.7910**** 1.90 (1)
SI → BI 0.6258**** 0.5865**** 0.7130**** -0.3379 1.47 (1)
FC → USE 0.3775*** 0.3683*** 0.3995** -0.4973*** 0.03 (1)

Notes:
For each potentially moderated effect, a model with measurement equivalence constraints and a parallel slopes assumption for
all latent variables was tested against the almost identical model for which the parallel slopes assumption was released for the
potentially moderated effect. Consequently, the likelihood ratio test that compares whether the models are significantly different
always has one degree of freedom.

Variables definition:
* ** *** ****
, , , indicate p<0.1, 0.05, 0.01, and 0.001 respectively

Source: Own illustration based on survey results


Table A.13: Analysis of the moderating effect of hierarchy level (3/3)
(Pot.) moderated Likelihood
Total sample Group samples
effect ratio test
n=265 Junior, n=194 Senior, n=71
Mean diff. [0-1] χ2 (df)
Unstd. coefficient Unstd. coefficient[0] Unstd. coefficient[1]
H1c: PE → BI 0.2361** 0.2236** 0.2521** -0.3436 0.08 (1)
H2c: EE → BI 0.0107 -0.0059 0.1060 -0.7386**** 1.83 (1)
SI → BI 0.6195**** 0.5804**** 0.7052**** -0.3794* 1.46 (1)
FC → USE1 0.3246*** - - - -

Notes:
For each potentially moderated effect, a model with measurement equivalence constraints and a parallel slopes assumption for
all latent variables was tested against the almost identical model for which the parallel slopes assumption was released for the
potentially moderated effect. Consequently, the likelihood ratio test that compares whether the models are significantly different
always has one degree of freedom.
1) The group model with a released parallel slopes assumption (and the minimization of its discrepancy function) could not be
calculated by Stata v15.1.

Variables definition:
* ** *** ****
, , , indicate p<0.1, 0.05, 0.01, and 0.001 respectively

Source: Own illustration based on survey results

Table A.14: Analysis of the moderating effect of voluntariness


(Pot.) moderated Likelihood
Total sample Group samples
effect ratio test
n=264 Low, n=115 High, n=149
Mean diff. [0-1] χ2 (df)
Unstd. coefficient Unstd. coefficient[0] Unstd. coefficient[1]
PE → BI 0.2292** 0.1601 0.2982*** -0.3196* 2.22 (1)
EE → BI -0.0036 -0.0355 0.0518 -0.2731 1.24 (1)
SI → BI1 0.6624**** - - - -
FC → USE 0.3214*** 0.4058*** 0.2373* -0.2019 1.10 (1)

Notes:
For each potentially moderated effect, a model with measurement equivalence constraints and a parallel slopes assumption for
all latent variables was tested against the almost identical model for which the parallel slopes assumption was released for the
potentially moderated effect. Consequently, the likelihood ratio test that compares whether the models are significantly different
always has one degree of freedom.
1) The group model with a released parallel slopes assumption (and the minimization of its discrepancy function) could not be
calculated by Stata v15.1.

Variables definition:
, , ,
* ** *** ****
indicate p<0.1, 0.05, 0.01, and 0.001 respectively

Source: Own illustration based on survey results


Table A.15: Analysis of the moderating effect of experience
(Pot.) moderated Likelihood
Total sample Group samples
effect ratio test
n=258 Low, n=139 High, n=119
Mean diff. [0-1] χ2 (df)
Unstd. coefficient Unstd. coefficient[0] Unstd. coefficient[1]
PE → BI 0.4031**** 0.4172**** 0.3704**** 1.1159**** 0.18 (1)
EE → BI -0.0296 0.0727 -0.0928 1.4855**** 4.05 (1)**
SI → BI 0.4892**** 0.5150**** 0.4345**** 0.8834**** 0.70 (1)
FC → USE 0.2124* 0.2547** 0.0967 1.1160**** 0.74 (1)

Notes:
For each potentially moderated effect, a model with measurement equivalence constraints and a parallel slopes assumption for
all latent variables was tested against the almost identical model for which the parallel slopes assumption was released for the
potentially moderated effect. Consequently, the likelihood ratio test that compares whether the models are significantly different
always has one degree of freedom.

Variables definition:
* ** *** ****
, , , indicate p<0.1, 0.05, 0.01, and 0.001 respectively

You might also like