Posts on Ray Yang, Ph.D.

Play the Thinking Game -- from Seeing from Eyes to Seeing with the Mind

Wed, 02 Jun 2021 00:00:00 +0000

The ontological and epistemological underpinning of research methodology.

Ontology: The truth lies in the things that we can’t see from our eyes. The invisible generates the visible under natural laws and processes. Thus, the “observational” data (what we see from our eyes) stems from the underlying mechanism (what we can’t see from our eyes). Then, our mind comes to help (to develop theories), with reasoning and inference, see what is behind the data.

Epistemology: We need to be suspicious about what we see from our eyes because data can be misguiding, especially when we shut off our reasoning mind and stop making the inference. In other words, the frontend phenomenon (e.g., the data) can be illusional when we stop seeking to understand the backend processes (e.g., the processes that generate the data). Researchers need to be self-reflective and self-critical about HOW we know what we know.

Inference: We can use what we see from our eyes (observational data) to recover the underlying truth with our mind (reasoning techniques). There are two elements: first, we know that what we see from eyes (e.g., data) comes from the underlying truth (e.g., probability); second, we know some properties about the truth (e.g., distribution) so that we generate estimates from the what our eyes see (e.g., sample).

Counterfactuals: The reality (i.e., a data point), events that have already occurred, is a probabilistic ‘realization’ of the underlying mechanism (i.e., probability distribution). To recover the truth, we need to construct the alternatives to the reality just like looking for the missing pieces to complete a puzzle. Fact (what happens) does not equate to the truth (which determines what happens). Fact itself does not predict fact. Truth, recovered from the fact and its alternative, will be able to predict fact.

Citation Analysis on Corporate Strategy & Network Change

Tue, 23 Apr 2019 00:00:00 +0000

1 Central Themes: Network of Co-occurrence Keywords

2 the “Classic” Papers

## [1] "most central papers (top 30) in co-citation network"

##       BAUM JAC 2000        ANJOS F 2015       GREVE H. 2014        LAVIE D 2006 
##            1.000000            1.000000            1.000000            1.000000 
##       PHELPS C 2012     PODOLNY JM 2001      PORRINI P 2004 TATARYNOWICZ A 2016 
##            1.000000            1.000000            1.000000            1.000000 
##       ZAHEER A 2010        AHUJA G 2000        ERDOS P 1959     GULATI R 1999-1 
##            1.000000            1.000000            1.000000            1.000000 
##  HASPESLAGH P. 1991    HERNANDEZ E 2015     HIGGINS MJ 2006         LIN N. 2001 
##            1.000000            1.000000            1.000000            1.000000 
##       ROGAN M. 2013        ROGAN M 2014        SYTCH M 2014       ZAHEER A 2005 
##            1.000000            1.000000            1.000000            1.000000 
##        BURT R. 1992      BUSKENS V 2008        DEVOS E 2009        AHUJA G 2012 
##            1.000000            1.000000            1.000000            1.000000 
##    BARABASI AL 1999       RYALL MD 2007      SHAVER JM 2006       ADLER N. 2005 
##            1.000000            1.000000            1.000000            0.572516 
##        ARORA A 1990         BALA V 2000 
##            0.572516            0.572516

Citation Analysis of Recent Acquisition Research

Mon, 15 Apr 2019 00:00:00 +0000

1 Sample and Data

The sample used in this analysis contains 160 articles on “mergers and acquisitions” published on top management and finance journals since 2008.

2 Central Themes

2.1 Network of Co-occurrence Keywords

This graph shows the most common themes (top 30 co-occuring keywords) of these papers.

2.2 MDS Mapping of the Conceptual Structure

We can use MDS (Multi-dimentional Scaling) and Dendrogram to map the distance/dissimilarity among the themes.

3 Central Papers

3.1 the “Main Stream” Papers

## [1] "most central papers (top 20) in bibliographic coupling network"

##   BARKEMA HG, 2008       KIM JY, 2009   NADOLSKA A, 2014 
##          1.0000000          0.8682856          0.8598660 
##       KIM JY, 2011 HALEBLIAN JJ, 2017   GORANOVA M, 2010 
##          0.8555986          0.8210859          0.7714827 
##    DEVERS CE, 2013 STEINBACH AL, 2017      KROLL M, 2008 
##          0.7253757          0.7121932          0.7104844 
##  MUEHLFELD K, 2012  EL-KHATIB R, 2015     ELLIS KM, 2011 
##          0.6669592          0.6612580          0.6551406 
##  MCDONALD ML, 2008      SHI W, 2017-1 HEIMERIKS KH, 2012 
##          0.6185902          0.5899796          0.5799630 
##  GRAEBNER ME, 2009      ZOLLO M, 2009     ELLIS KM, 2009 
##          0.5753946          0.5740604          0.5715228 
##        YIM S, 2013      ZOLLO M, 2010 
##          0.5578773          0.5558791

To identify the “main stream” papers, we can construct a “bibliographic coupling” (BC) network to see which papers co-cited the same prior research with other papers. The top 20 “main stream” (by eigenvector centrality in the BC network) papers.

3.2 the “Classic” Papers

## [1] "most central papers (top 20) in co-citation network"

##    MOELLER SB 2004   HALEBLIAN J 1999     JENSEN MC 1986 
##          1.0000000          0.9891969          0.9313744 
##       MORCK R 1990   HAYWARD MLA 2002 HASPESLAGH P. 1991 
##          0.8737119          0.8477468          0.7896179 
##   HAYWARD MLA 1997       KING DR 2004        ROLL R 1986 
##          0.7119846          0.7104482          0.6958790 
##       ZOLLO M 2004    MASULIS RW 2007  MOELLER SB 2005-1 
##          0.6931354          0.6809497          0.6665792 
##     ASQUITH P 1983     ANDRADE G 2001     JENSEN MC 1976 
##          0.6269530          0.6209498          0.5782434 
##  MALMENDIER U 2008  HALEBLIAN JJ 2006      FULLER K 2002 
##          0.5702609          0.5595749          0.5314969 
##  HECKMAN JJ 1979-1     GOMPERS P 2003 
##          0.5272799          0.5248858

To identify the “classic” papers, we can construct a “co-citation” (CC) network to see which papers are cited together with other papers by later researchers. The top 20 “classic” (by eigenvector centrality in the CC network) papers.

A.I./Blockchain Adoption and Technology Substitution

Tue, 02 Apr 2019 00:00:00 +0000

Artificial intelligence (A.I.) and blockchain are the hottest new technologies right now. It would be fun to think about the future of them (see the patterns, draw the inference, and make predictions). Would A.I. and blockchain be the next “internet,” which has disrupted many traditional business models and creates many new giants across the globe? Or, since businesses have learned a lesson from the internet disruption, A.I. and blockchain will be merely incorporated into the existing technology terrain as an expansion of opportunities by the incumbents? Do A.I. and blockchain have the same fate?

I’ve been leaving an eye on the adoption/acquisition of new technology in the banking sector (since one of my dissertation chapters was on banking acquisitions). Here is some latest news.

The banking sector has a bright prospect incorporating the A.I. in its six core functions.

JP Morgan has adopted cryptocurrency in the payment businesses.

Wells Fargo and Mastercard CEOs express doubt about blockchain.

PayPal made a major investment on the blockchain.

Could there be a unified explanation for what is happening right now? Well, one may associate the topic of technology development with Clayton Christensen’s ‘innovator’s dilemma’ assertion, which predicts emerging new technology will quickly take over declining old ones (often depicted as the double-S curves). It’s the good management–trying best to hear about customer needs, play with competition carefully, and align resource allocation with calculated risk/benefit profiles–who prevent the incumbents from adopting new technologies. Then shouldn’t managers try their best? According to Clayton Christensen, it is difficult for good management not to do good–it is difficult for the incumbents to utilize new technologies even they tend to spot them earlier than startups. At the initiation stage of new technology, incumbents can’t risk their existing market share and customer relationships with unproven models, whereas startups have little to lose. The prediction is: startups disrupt incumbents.

However, things are different right now. The boundary between new tech. and old tech. are melting down; they are intertwined ecosystems rather than isolated communities. Incumbents right now have taken deliberate effort and allocate resource for emerging technology. They are agiler and more determined to disrupt themselves before being disrupted by others. If important pre-existing conditions are different, it might be necessary to modify the predictions from Clayton Christensen’s double-S curves .

First, the compatibility between incumbents’ existing technology and emerging technology should lower the likelihood and rate of disruption. Let’s consider two cases here. One is about different incumbent technologies: as compared to traditional banking, the higher level of digitization of PayPal allows a higher compatibility potential with blockchain (consider internet and blockchain don’t rely on a central authority/agency as traditional banking does). The other is about different emerging technologies: as compared to the blockchain, A.I. allows a higher level of compatibility with traditional banking (both banking and A.I. are aimed at decision/prediction accuracy and efficiency.)

Second, the intensity of the incumbent technology expansion should also lower the likelihood of disruption. This factor is not only related to incumbents’ existing resource base but also has something to do with management. A recent paper in SMJ ¹ has updated the “double-S” framework by incorporating the “incumbent technology expansion” in predicting the rate of substitution and found empirical support for the model’s prediction efficacy.

Conclusion: since A.I. tends to have a higher level of compatibility with, and also face a stronger expansion threat from, the incumbent technology than blockchain, it’s predicted to have a lower likelihood/rate of disruption. However, contingencies should also exist due to the significant variance in incumbents’ technology, resource, and management conditions.

Adner, R., & Kapoor, R. (2016). Innovation ecosystems and the pace of substitution: Re‐examining technology S‐curves. Strategic management journal, 37(4), 625-648.↩

A Summary of Recent 'Social Evaluation' Research

Mon, 25 Mar 2019 00:00:00 +0000

1 Motivation

The goal of this post is to analyze and review the “social evaluation” research tracing back to Zuckerman (1999)¹. The analysis/review is based on hand-collected papers that directly and indirectly cite Zuckerman (1999). The review is not intended to be comprehensive but to understand the central ideas on the topic developed by management/sociology scholars in the past 20 years.

Let’s continue on the “positioning” topic from one of my previous posts (firms are social actors being evaluated by social audiences), but shift our focus of discussion from “what actors do” to “what audience think.” That is, instead of an “actor” standpoint, let’s take an “audience” perspective to consider how firms as market participants are evaluated by market audiences. Reality doesn’t directly translate into human decisions; human beings can only understand reality from their brain, which is subjected to social construction. As a result, the evaluations on market participants do not come from complete rationality in a social vacuum. Rather, audiences’ cognition is shaped by the market’s categorical structure and/or the socially accepted categorical beliefs.

2. Biblio-Analysis

2.1 Selection Criteria

## Warning: The curvature argument has been deprecated in favour of strength

The raw data is based on 29 hand-collected papers directed and indirectly citing Zukerman (1999) and extracted from Web of Science. The papers that not on the topic of “social evaluation” are excluded from the sample, based on the author’s judgment. We can get a general idea of the citation relationship among those papers from the direct citation network graph.

2.2 Central Sources and Keywords

By quickly checking the origins (papers being cited) of the sampled papers, we can see that a big portion (the blue-colored cluster) is the classic institution work. It makes sense, as the discussion of the underlying mechanism of social evaluation should be related to how the evaluating standards are formed in the institutionalization processes. The summary in 3.2 will be based on the top-cited “social evaluation” papers in the red-colored cluster.

The inter-connected keyword clusters present the relationship among the main topics covered by this stream of research. It seems to have a quite strong “strategy” flavor². The most frequently used keywords include “industry”, “market”, and “performance.”

3 Classic Papers

3.1 Top Citations

by frequency:

## [1] "ZUCKERMAN EW, 1999, AM J SOCIOL, V104, P1398, DOI 10.1086/210178"          
## [2] "HSU G, 2006, ADMIN SCI QUART, V51, P420, DOI 10.2189/ASQU.51.3.420"        
## [3] "HSU G, 2009, AM SOCIOL REV, V74, P150, DOI 10.1177/000312240907400108"     
## [4] "PONTIKES EG, 2012, ADMIN SCI QUART, V57, P81, DOI 10.1177/0001839212446689"
## [5] "RAO H, 2005, AM SOCIOL REV, V70, P968, DOI 10.1177/000312240507000605"

by co-citation centrality:

## [1] "most central sources (top 5)"

## ZUCKERMAN EW 1999        HSU G 2009      HSU G 2006-1  PONTIKES EG 2012 
##         1.0000000         0.7142248         0.6970316         0.6754593 
##        RAO H 2005 
##         0.6150981

The most frequent (by frequency count) citations are also the most central (by eigenvector centrality) citations.

3.2 Summary of Top-cited Papers

Hsu et al. (2009) provide an integrative explanation for the multi-category disadvantage. The paper combines the actor- and audience-centered explantions and argues that category spanning is both difficult for the actor to manage and hard for the audience to understand. Hsu (2006) explicitly analyzes the underlying cognition mechanism in which film audiences encounter difficulties in making sense of multi-genre films. Pontikes (2012) analyzes the different segments of market audiences with distinct preferences for ambiguous classification–the difficulty of classifying startups into established categories. According to her, consumers and corporate venture capitalists are more passive market-takers who prefer startups that are easy to classify; in contrast, individual venture capitalists are market-makers who have the power to redefine the market structure and see ambiguous classification as opportunities. Rao et al. (2005) argue that the market boundaries are melting when high-status market participants’ cross the market categories–French chefs with starts in the Guide Michelin make hybrid cuisines. Typically, there are cost and risk associated with “cross-category borrowings.” But they tend to attenuate when borrowings prevail. Overall, it seems to be risky to cross the categorical boundaries. But contextual contingencies for the benefit/risk profile of category-straddling exists.

References:

Zuckerman, E. W. (1999). The categorical imperative: Securities analysts and the illegitimacy discount. American Journal of Sociology, 104(5), 1398-1438.

Hsu, G., Hannan, M. T., & Koçak, Ö. (2009). Multiple category memberships in markets: An integrative theory and two empirical tests. American Sociological Review, 74(1), 150-169.

Hsu, G. (2006). Jacks of all trades and masters of none: Audiences’ reactions to spanning genres in feature film production. Administrative Science Quarterly, 51(3), 420-450.

Pontikes, E. G. (2012). Two sides of the same coin: How ambiguous classification affects multiple audiences’ evaluations. Administrative Science Quarterly, 57(1), 81-118.

Rao, H., Monin, P., & Durand, R. (2005). Border crossing: Bricolage and the erosion of categorical boundaries in French gastronomy. American Sociological Review, 70(6), 968-991.

The central idea of this paper is that firms’ stock price is affected by the categorical pattern of financial analysts’ industry coverage, which reflects the institionalized expectations for firms’ product mix straddling across various market categories.↩
For the central topics in strategy research, refer to “the intellectual structure of the strategic management field…” by Nerur et al. (2008) and “what is strategic management…” by Nag el al. (2007)↩

Interoperability in Tech. Evolution—Thoughts on “Reticulate"

Mon, 11 Mar 2019 00:00:00 +0000

“In our highly dynamic world, it’s not enough for an organization to possess a competitive advantage at a point in time; it needs an evolutionary advantage over time—a capacity to change as fast as change itself; to change before a crisis breaks.” — Gary Hamel

On March 6th, the R package “reticulate”(1.11.1) was released on Cran, which allows R users to directly call Python objects or use Python in an R interface.¹ Bloggers have been praising this development for the ground-breaking interoperability between R and Python. With “reticulate,” we no longer need to jump between different IDEs to leverage the complementarity between R and Python. This will certainly bring big ease to the programming routine of many “bilingual” data scientist using both R and Python.

As an occupational habit, I tend to think about the value creation and value capture in the dynamics of industry and technology evolution. Here are some thoughts on the “reticulate” case.

First of all, one the user level, “reticulate” by Rstudio will benefit those who program in both R and Python. So far, R and Python are the two most popular programming languages in data science. A decision for a newcomer in data science to make is always: “R vs. Python, which one should I choose?” Now this headache is significantly lessened. If we can overcome the “language barrier” easily, we won’t be stuck in the either-or any more.

Second, on the community level, there have already been quite a few early adopters who are R-Python bilinguals even before the pre-release of “reticulate.” They choose a language depending on which one is more readily usable for a specific problem. For bilinguals, there is no transition in their head, but there are some inefficient transitions on their fingertips. Now, they will leverage the power of “interoperability” to keep their hand movements in one place–Rstudio. These users will be engaged in Rstudio more often. What will also naturally happen? (a) the total number of R-Python bilinguals will grow; (b) the two languages will become more integrated, instead of being divided. We will see fewer (or slower growth of) packages/modules which have the same name and perform the same functionality in two languages, and also less “R vs Python comparisons.”

Third, thinking from the technological (and also strategic) perspective, we see the greatly increased compatibility between two programming languages. What follows compatibility will be an increase in the adoption rate on the technology who opens up its door to embrace other technologies. Rstudio is the superpower in the R domain. Such a move will draw many Python programmers to R and also attract and retain R users with the empowerment of R-Python interoperability. This “one-way” compatibility² will greatly enhance the usability and desirability of the Rstudio as a platform provider.

The release of “reticulate” is a strategic move of Rstudio. More data scientists will be stay engaged in Rstudio, or they will spend more fo their time on the platform. But this is not the end of the story. The interoperability will continuously enlarge the overlap between the R and Python communities if we draw a Venn diagram in our mind. The Python-centered IEDs are also powerful. They have great features that their users wouldn’t live without. They may match the move of Rstudio. More likely, the reverse “recuticate” in Python will emerge from the large open source community, by innumerable open source developers. Even if the one-way compatibility will continue, Python IEDs will not lose the game. As R programmers can easily operate Python in Rstudio, they are also more likely to travel and contribute to the Python world. This is similar to the case when Apple allows Microsoft’s Windows operating system to be installed on Mac and Amazon’s Kindle electronic book app installed on iPad. The competing platforms will both benefit from the one-way compatibility, because the enlarged overlap of the Venn diagram will bring “two-way” flows of traffic to both sides.³

Open source can always provide more flexibility to the users, developers, and businesses involved. Perhaps we cannot formulate the dynamics of technological evolution in a game-theoretic fashion, as the players in open source are not “competing” in the same way. For example, Anaconda, the company who distributes many Python IDEs (including JupyterLab and Spyder), has long been providing Rstudio IDE on their cloud and GUI. I see “reticulate” as a strategic, but also very harmonious move to create net benefits to the open source community for data science at large.

This can be easily done in Rstudio. Webinar by Rstudio Documentation by Rstudio ↩︎
The existing python IDEs have not provided an interface for R-Python interoperability yet. Python programmers tend to use JupyterLab, Rodeo, Spyder, Visual Studio Code, and PyCharm, none of which have such an interoperability feature right now.↩︎
For a more detailed analysis: Adner, R., Chen, J., & Zhu, F. (2016). Frenemies in Platform Markets: The Case of Apple’s iPad vs. Amazon’s Kindle. Harvard Business School Technology & Operations Mgt. Unit Working Paper, (15-087).↩︎

“MMeM”: Modeling the Multivariate Mixed-effects

Sun, 10 Mar 2019 00:00:00 +0000

I co-developed an R pakcage “MMeM” for estimating the variance-covariance matrix of random effects \(\mathbf{u}\) and \(\mathbf{e}\) on multiple dependent variables.

Statistical Intuition

The univariate mixed-effects model

In univariate mixed-effects model: \(\mathbf{y} = \mathbf{Xb} + \mathbf{Zu} + \mathbf{e}\) (1), we estimate the variance component \(\sigma_u^2\) and \(\sigma_e^2\) for \(\mathbf{u} \sim N(\mathbf{0}, \sigma_u^2\mathbf{I})\) and \(\mathbf{e} \sim N(\mathbf{0}, \sigma_e^2\mathbf{I})\).

In formula (1): \(\mathbf{y}\) is n \(\times\) 1 response vector; \(\mathbf{X}\) and \(\mathbf{Z}\) are n \(\times\) p and n \(\times\) s; \(\mathbf{b}\) is p \(\times\) 1 coefficients vector for the fixed effects; \(\mathbf{u}\) is s \(\times\) 1 matrix for the random effects, \(\mathbf{e}\) is n \(\times\) 1 vector of random errors.

The multivariate mixed-effects model

In multivariate mixed-effects model: \(\mathbf{y} = (\mathbf{I} \otimes \mathbf{X})\mathbf{b} + (\mathbf{I} \otimes \mathbf{Z} )\mathbf{u} + \mathbf{e}\) (2), in which \(\mathbf{y} = \left\{\mathbf{y}_i\right\}_c, \mathbf{b} = \left\{\mathbf{b}_i\right\}_c, \mathbf{u} = \left\{\mathbf{u}_i\right\}_c, \mathbf{e} = \left\{\mathbf{e}_i\right\}_c, i =1, \dots, q\), we estimate the variance-covariance matrix of random effects \(\mathbf{u}\) and \(\mathbf{e}\) on \(q\) response variates, namely \(\mathbf{T}\) and \(\mathbf{E}\), for \(var(\mathbf{u}) = \mathbf{G} = \mathbf{T}\otimes \mathbf{I}_s, var(\mathbf{e}) = \mathbf{R} = \mathbf{E} \otimes \mathbf{I}_n, var(\mathbf{y}) = \mathbf{V} = \mathbf{T}\otimes \mathbf{ZZ}' + \mathbf{E} \otimes \mathbf{I}_n\).

In formula (2): \(\mathbf{y}\) is nq \(\times\) 1 response vector; \(\mathbf{X}\) is n \(\times\) p design matrix for the fixed effects; \(\mathbf{b}\) is pq \(\times\) 1 coefficients vector for the fixed effects; \(\mathbf{Z}\) is n \(\times\) s design matrix for the random effects; \(\mathbf{u}\) is sq \(\times\) 1 vector of the random effects; \(\mathbf{e}\) is nq \(\times\) 1 vector of random errors.

An Implementation Example

# install.packages("MMeM")
library(MMeM)
data(simdata)

Initialization

# initialize with a positive-definiate var-cov
T.start = matrix(c(10, 5, 5, 15), 2, 2)
E.start = matrix(c(10, 1, 1, 3), 2, 2)

Estimation

# using the Henderson3 estimation mothod
results_henderson = MMeM_henderson3(fml = c(V1,V2) ~ X_vec + (1|Z_vec), data = simdata, factor_X = TRUE)

## Bivariate response: V1 and V2

print(results_henderson)

## $T.estimates
##                 T: V1   T: V1 V2     T: V2
## T.estimates  65.47395   9.969188  9.766204
## T.variance  807.56303 128.824836 20.552160
## 
## $E.estimates
##                E: V1  E: V1 V2     E: V2
## E.estimates 55.74506 11.477502 41.171927
## E.variance  11.09826  8.196899  6.054027

“regrrr”: One-stop R-toolkit for Compiling Regression Results

Wed, 06 Mar 2019 00:00:00 +0000

In strategy/management research, we always need to compile the regression results into the publishable format and sometimes plot the moderating effects. Thus, I developed this “regrrr” package to help do the job.

Here is the quickstart guide.

Installation

To install from CRAN:

install.packages("regrrr")
library(regrrr)

You can also use devtools to install the latest development version:

devtools::install_github("raykyang/regrrr")
library(regrrr)

Examples

compile the correlation table

library(regrrr)

## Registered S3 methods overwritten by 'tibble':
##   method     from  
##   format.tbl pillar
##   print.tbl  pillar

data(mtcars)
m0 <- lm(mpg ~ vs + carb + hp + wt, data = mtcars)
m1 <- update(m0, . ~ . + wt * hp)
m2 <- update(m1, . ~ . + wt * vs)
cor.table(data = m2$model)

##          Mean  S.D.     1     2    3    4    5
## 1.mpg   20.09  6.03  1.00                     
## 2.vs     0.44  0.50  0.66  1.00               
## 3.carb   2.81  1.62 -0.55 -0.57 1.00          
## 4.hp   146.69 68.56 -0.78 -0.72 0.75 1.00     
## 5.wt     3.22  0.98 -0.87 -0.55 0.43 0.66 1.00

compile the regression table

regression_table <- rbind(
combine_long_tab(to_long_tab(summary(m0)$coef),
                 to_long_tab(summary(m1)$coef),
                 to_long_tab(summary(m2)$coef)),
compare_models(m0, m1, m2))
rownames(regression_table) <- NULL
print(regression_table)

##        Variables   Model 0   Model 1   Model 2
## 1    (Intercept) 35.435*** 48.157*** 46.698***
## 2                  (2.503)   (4.097)   (9.272)
## 3             vs     1.353     1.077     2.171
## 4                  (1.382)   (1.152)   (6.320)
## 5           carb    -0.057    -0.043    -0.009
## 6                  (0.449)   (0.374)   (0.426)
## 7             hp   -0.024† -0.113***   -0.107*
## 8                  (0.014)   (0.027)   (0.044)
## 9             wt -3.792*** -8.071***   -7.594*
## 10                 (0.658)   (1.307)   (3.016)
## 11         hp:wt             0.027**    0.025†
## 12                           (0.008)   (0.015)
## 13         vs:wt                        -0.367
## 14                                     (2.081)
## 15     R_squared     0.833     0.889     0.889
## 16 Adj_R_squared     0.808     0.867     0.862
## 17       Delta_F            12.517**     0.031

plot the moderating effect

plot_effect(reg.coef = summary(m2)$coefficients, data = mtcars, model = m2,
            x_var.name = "wt", y_var.name = "mpg", moderator.name = "hp",
            confidence_interval = TRUE,  CI_Ribbon = FALSE, 
            xlab = "Weight", ylab = "MPG", moderator.lab = "Horsepower") +
ggplot2::theme(text=ggplot2::element_text(family="Times New Roman", size = 16))

plot the moderating effect with a linear spline

library(lspline)
data(mtcars)
m3 <- lm(mpg ~ vs + carb + hp + lspline(wt, knots = 4, marginal = FALSE) * hp, data = mtcars)
plot_effect(reg.coef=summary(m3)$coefficients, data = mtcars, model = m3, 
            x_var.name = "wt", y_var.name = "mpg", moderator.name = "hp",
            xlab="Weight", ylab="MPG", moderator.lab="Horsepower") +
ggplot2::theme(text=ggplot2::element_text(family="Times New Roman", size = 16))

As we can see from the last line of code, the plot is customizable using “ggplot2”. There are a couple of other functions. Please see the reference manual on R documentation for details.

I’m also aiming to expand the package’s usage around its core functions. If you have any ideas or want to report a bug, please contact me or suggest on the GitHub page.

A Quick Bibliometric Analysis on “Positioning”

Tue, 26 Feb 2019 00:00:00 +0000

I had a whim and wanted to do some quick bibliometric analysis on the topic of “strategic positioning”.

The raw data is extracted from Web of Science. The search parameters:

TITLE: ("strategic positioning" OR "industry positioning" OR "market positioning")
Refined by: WEB OF SCIENCE CATEGORIES: ( MANAGEMENT OR BUSINESS OR ECONOMICS ) AND WEB OF SCIENCE CATEGORIES: ( MANAGEMENT OR BUSINESS OR INTERNATIONAL RELATIONS OR ECONOMICS ) AND DOCUMENT TYPES: ( ARTICLE ) AND LANGUAGES: ( ENGLISH )
Timespan: All years. 
Indexes: SCI-EXPANDED, SSCI, A&HCI, CPCI-S, CPCI-SSH, BKCI-S, BKCI-SSH, ESCI, CCR-EXPANDED, IC.

I used the R package “bibliometrix” to conduct (1) “bibliographic coupling” and (2) “co-citation” analysis.

Bibliographic coupling analysis allows us to explore the “confluence and interactions” among recent papers. Two articles are bibliographically coupled if they co-cited at least one article.

Co-citation analysis allows us to trace back to the origin of thoughts and classic papers. Two articles have a co-citation tie if they are both cited in a third article.

Reference:

Aria, M. & Cuccurullo, C. (2017) bibliometrix: An R-tool for comprehensive science mapping analysis, Journal of Informetrics, 11(4), pp 959-975, Elsevier.

Optimal Positioning for Firms and Individuals

Mon, 25 Feb 2019 00:00:00 +0000

Imagine you are invited to a party. You want to get noticed by other partygoers. Or you want to be recognized as a “cool” person by those who you are interested in. The “coolness” will result in some social benefits you desire. This is a scene of “social evaluation,” in which you want to stand out of the crowd of peer actors (competitors) to impress the audiences (customers and suppliers). If you succeed, the peer actors will admire you, the audiences will adore you, and of course, the esteems and profitable relationships will follow. Now the question is, “how to position yourself to be ‘cool’ so that you can win the game?”

Practically, the first question coming to your mind might be “to what extent should I be similar to or different from others?” You want to be different so that you can stand out. However, you don’t want to be too different so that people think you are wired. Your goal is to “impress” but not to “surprise.” It is a bad idea to violate the audience’s expectations on your appearances and behaviors – the commonly accepted “norms.” Violations typically lead to penalties.

Excessive differentiation is risky. It jeopardizes your “legitimacy,” the potential for being accepted by the audience, who controls the resources and opportunities. Now the headache you get is — being similar places you under competitive pressures (so you can’t stand out) but being different gives you the risk of losing legitimacy. Then, you may find a simple solution: find a middle ground between the two extremes — “to be different as ‘legitimately’ possible.” Good. You got a “holistic” approach, in which you consider all the audiences, attributes, and evaluation processes as a “whole.” And you have a clear goal—maximizing the total audience perceptions. It is implementable. It is the wisdom of Confucius’ as well as Aristotle’s “Golden Mean.”

If you don’t want to stop here, let’s think further. We can do some analysis¹, by moving away from the “holistic” approach to an “analytical” one. For example, we can anatomize the positioning problem in three ways: (1) separating the evaluating audiences, (2) separating the actor attributes, and (3) separating the evaluation processes.

First, different audience groups (government, customers, investors, etc.) may demand different levels of conformity vs. differentiation, so you can deploy specific positioning strategy to target specific audiences (e.g. horizontal websites tend to give different people different things while vertical websites tend to meet specific people’s specific needs). Second, as audiences allocate their decision weights among multiple dimensions of difference, you can conform on most of them to maintain legitimacy and then differentiate on a couple of others to look cool (e.g. in a party, people always care about dress code while worrying about outfit clash). Third, as the audiences typically engage in a categorization process before making the final selection, you can strategize for “entering the game” and “winning the game” separately (e.g. audiences tend to make a short list first, select the winner from it. No matter you want to design a website or find a job, to make it to the shortlist, you need to first figure out the similarities that help you stay relevant; and in the selection stage, find the differences that help you win).

The positioning problem shouldn’t stop here. As audiences evaluate actor attributes with their subjective minds, numerous factors (cognitive, sociological, cultural ones) determine how individual opinions aggregate and emerge as market behavior. Researchers have been studying the patterns, drawing inferences and making predictions. Importantly, there could exist positive and negative feedback loops for the mutual influence between actor behavior (strategic action) and audience perception (performance evaluation). This is what I’m interested in.

References:

Deephouse, D. L. 1999. To be different, or to be the same? It’s a question (and theory) of strategic balance. Strategic Management Journal, 20(2): 147–166.

Zuckerman, E. W. 2016. Optimal distinctiveness revisited: An integrative framework for understanding the balance between differentiation and conformity in individual and organizational identities. Handbook of Organizational Identity.

Zhao, E. Y., Fisher, G., Lounsbury, M., & Miller, D. 2017. Optimal distinctiveness: Broadening the interface between institutional theory and strategic management. Strategic Management Journal, 38(1): 93–113.

Merriam-Webster has a definition of “analysis” — “separation of a whole into its component parts.”↩︎