Common-method bias

There exist two different techniques to control for common methods bias Choice of study design and statistical procedures (Podsakoff et al. 2003). 

The statistical controls for common method bias were only carried out for the full (mediated) model, as the indicators for all constructs were obtained from the same source. Three statistical remedies were used First, Harman s single factor test was employed. 1 applied factor analysis without rotation to check for the fit of a single factor model. The variance extracted for a one factor solution was below 50% (35,05%). With eigenvalues >1, six factors were extracted. Thus one can conclude that there is no single... 

Figure 11 PLS model for assessing common method bias This procedure has been first suggested by Liang et al. (2007), and has also been employed by others (e.g. Herath and Rao 2009). Use experience is originally a formative constructs, but has to be specified reflectively in order to being able to interpret the results for this test I compared two versions of the theoretic model, which only differed in the specification of the latent construct use experience (formative vs. reflective) (cf. Herath and Rao 2009 Liang et al. 2007). There were no qualitative differences between the two models. Thus both indicators of use experience were specified as reflective...

Even so, I designed my study to minimize this potential risk Unless indicated otherwise, my precautions follow Podsakoff et al. (2003] The questionnaire was checked by the person responsible for disseminating the survey in each Arm for its unequivocality. In the introduction, anonymity was assured to all participants. I posed specific questions about product use and behavior in the firm, rather than about abstract attitudes (Cote and Buckley 1987], Questions relating to predictor and criterion variables were intermingled rather than sequential and used different types of scales (Likert scales, ordinal scales, semantic differentials]. If possible, I excluded reverse coded items as they may favor common method bias (DiStefano and Motl 2006]. 

I also performed statistical tests for common method bias on firms 1.2. and 3 (n=83]. (The same was not possible for firm 4 due to the lack of data on the antecedents of lead... 

First, I conducted Harman s single factor test to see if a single unrotated factor solution accounted for most of the variance (Malhotra et al. 2006 Podsakoff et al. 2003]. From this analysis of ali measured items, 10 factors emerged (eigenvalues >1 were extracted] the first factor explained a variance of 32,3%. Thus, according to Harman s test, there is no severe common method bias. 

Second, I checked the bivariate correlations of the constructs for vaiues above 0,9 that could indicate common method bias (cf. Pavlou et al. 2007], Fortunately, my hlg]iest value was r=0,688 (see Table 25]. 

Third, 1 modeled an unobserved latent conunon methods foctor on wdiich all reflective items load, a procedure proposed by Uang et aL (2007] and Podsakoff et aL (2003]. This factor explains all the variance that caimot be attributed to the substantiated relationships or to error the remaining variance can then be attributed to common method bias. To simultaneously estimate all relationships, and in view of reduced... 

The accuracy of an analysis can be determined by several procedures. One common method is to analyze a known sample, such as a standard solution or a quality control check standard solution that may be available commercially, or a laboratory-prepared standard solution made from a neat compound, and to compare the test results with the true values (values expected theoretically). Such samples must be subjected to all analytical steps, including sample extraction, digestion, or concentration, similar to regular samples. Alternatively, accuracy may be estimated from the recovery of a known standard solution spiked or added into the sample in which a known amount of the same substance that is to be tested is added to an aliquot of the sample, usually as a solution, prior to the analysis. The concentration of the analyte in the spiked solution of the sample is then measured. The percent spike recovery is then calculated. A correction for the bias in the analytical procedure can then be made, based on the percent spike recovery. However, in most routine analysis such bias correction is not required. Percent spike recovery may then be calculated as follows ... 

Several types of bias are common in analytical methodology, including laboratory bias and method bias. Laboratory bias can occur in specific laboratories, due to an uncalibrated balance or contaminated water supply, for example. This source of bias is discovered when results of interlaboratory studies are compared and statistically evaluated. Method bias is not readily distinguishable between laboratories following a standard procedure, but can be identified when reference materials are used to compare the accuracy of different methods. The use of interlaboratory studies and reference materials allows experimentalists to evaluate the accuracy of their analysis. 

The depletion layer profile contains information about the density of states distribution and the built-in potential. The depletion layer width reduces to zero at a forward bias equal to and increases in reverse bias. The voltage dependence of the jimction capacitance is a common method of measuring W V). Eq. (9.9) applies to a semiconductor with a discrete donor level, and 1 is obtained from the intercept of a plot of 1/C versus voltage. The 1/C plot is not linear for a-Si H because of the continuous distribution of gap states-an example is shown in Fig. 4.16. The alternative expression, Eq. (9.10), is also not an accurate fit, but nevertheless the data can be extrapolated reasonably well to give the built-in potential. The main limitation of the capacitance measurement is that the bulk of the sample must be conducting, so that the measurement is difficult for undoped a-Si H. 

Traditionally, electrokinetic injection is the most common method for sample injection and separation into the microchannel utilizing some form of electroosmotic pumping. However, depending on the sample, due to the different mobility in an electric field, a bias to anionic, neutral, or cationic analytes is possible. Careful adjustment of pH and ionic strength can reduce this effect [2]. A broad injection band and sample leakage phenomena are also important defects of the electrokinetic injection method. Both are known to reduce the separation efficiency of a device since they result in a wide sample distribution within the microchannel and an increasing signal baseline as the number of injection runs increases [3-6], respectively. Other injection methods, like pressure injection, do not show this method-dependent effect. 

There are four major types of sampling methods random, stratified, systematic, and cluster. Random is by far the most commonly employed method in toxicology. It stresses the fulfillment of the assumption of avoiding bias. When the entire pool of possibilities is mixed or randomized (procedures for randomization are presented in a later section), then the members of the group are selected in the order that are drawn from the pool. 

---