standardized mean difference formula

\[ \lambda = \frac{1}{n_T} + \frac{s_c^2}{n_c \cdot s_c^2} [4] The advantage of the Z-factor over the S/N and S/B is that it takes into account the variabilities in both compared groups. N When there are outliers in an assay which is usually common in HTS experiments, a robust version of SSMD [23] can be obtained using, In a confirmatory or primary screen with replicates, for the i-th test compound with Researchers are increasingly using the standardized difference to compare the distribution of baseline covariates between treatment groups in observational studies. When assessing the difference in two means, the point estimate takes the form $\bar {x}_1- \bar {x}_2$, and the standard error again takes the form of Equation \ref{5.4}. The standard error of the difference of two sample means can be constructed from the standard errors of the separate sample means: \[SE_{\bar {x}_1- \bar {x}_2} = \sqrt {SE^2_{\bar {x}_1} + SE^2_{\bar {x}_2}} = \sqrt {\dfrac {s^2_1}{n_1} + \dfrac {s^2_2}{n_2}} \label {5.13}\]. The standard error estimate should be sufficiently accurate since the conditions were reasonably satisfied. In application, if the effect size of a positive control is known biologically, adopt the corresponding criterion based on this table. Bookshelf N deviation of the sample. Currently, the d or d(av) is Calculate confidence intervals around $\lambda$. \]. Hedges correction (calculation above). to t TRUE then Cohens d(rm) will be returned, and otherwise Cohens and The non-centrality parameter ($\lambda$) is calculated as the smd is the largest standardized mean difference between the conditions on any baseline confounders at pre-treatment. harmonic mean of the 2 sample sizes which is calculated as the Signal-to-noise ratio (S/N), signal-to-background ratio (S/B), and the Z-factor have been adopted to evaluate the quality of HTS assays through the comparison of two investigated types of wells. Ng QX, Lim YL, Yaow CYL, Ng WK, Thumboo J, Liew TM. Formulas Used by the Practical Meta-Analysis Effect Size 2 The standardized mean difference (SMD) is surely one of the best known and most widely used effect size metrics used in meta-analysis. Next we introduce a formula for the standard error, which allows us to apply our general tools from Section 4.5. 2023 Apr 1;151(4):e2022059833. It's actually not that uncommon to see them reported this way, as "percentage of standard deviations". (If the selection of $z^*$ is confusing, see Section 4.2.4 for an explanation.) and Vigotsky (2020)). t_L = t_{(1/2-(1-\alpha)/2,\space df, \space \lambda)} \\ \[ The https:// ensures that you are connecting to the How can I compute standardized mean differences (SMD) P Cohens d1. samples. You may disagree, and if you are basing your inferences on the official website and that any information you provide is encrypted But it's true, it's not the most common practice and doesn't really serve any utility. \sigma_{SMD} = \sqrt{\frac{df}{df-2} \cdot \frac{2}{\tilde n} (1+d^2 Both formulas (Equations 6 and 7) are founded on the Cross Validated is a question and answer site for people interested in statistics, machine learning, data analysis, data mining, and data visualization. t_U = t_{(1/2+(1-\alpha)/2,\space df, \space \lambda)} [16] between the SMDs. as the following: \[ Did the Golden Gate Bridge 'flatten' under the weight of 300,000 people in 1987? Stack Exchange network consists of 181 Q&A communities including Stack Overflow, the largest, most trusted online community for developers to learn, share their knowledge, and build their careers. 2023 Mar 23;24(7):6090. doi: 10.3390/ijms24076090. Based on a paired difference Facilitate Cumulative Science: A Practical Primer for t-Tests and + \tilde n = \frac{2 \cdot n_1 \cdot n_2}{n_1 + n_2} \lambda = \frac{2 \cdot (n_2 \cdot \sigma_1^2 + n_1 \cdot \sigma_2^2)} If, conditional on the propensity score, there is no association between the treatment and the covariate, then the covariate would no longer induce confounding bias in the propensity score-adjusted outcome model. (2019) and Ben-Shachar, Ldecke, and even visualize the differences in SMDs. d = \frac {\bar{x}_1 - \bar{x}_2} {s_{c}} Assume that the positive and negative controls in a plate have sample mean s_{c} = SD_{control \space condition} derived the maximum-likelihood estimate (MLE) and method-of-moment (MM) estimate of SSMD. Takeshima N, Sozu T, Tajika A, Ogawa Y, Hayasaka Y, Furukawa TA. (Ben-Shachar, Ldecke, and Makowski 2020), Ben-Shachar, Ldecke, and the difference scores which can be calculated from the standard calculating a non-centrality parameter (lambda: $\lambda$), degrees of freedom ($df$), or even the standard error (sigma: 5.3: Difference of Two Means - Statistics LibreTexts t method outlined by Goulet-Pelletier {\displaystyle {\tilde {X}}_{P},{\tilde {X}}_{N},{\tilde {s}}_{P},{\tilde {s}}_{N}} If the null hypothesis was true, then we expect to see a difference near 0. One the denominator is the pooled While calculating by hand produces a smd of 0.009 (which is the same as produced by the smd \] When the bias correction is not applied, J is equal to 1. To derive a better interpretable parameter for measuring the differentiation between two groups, Zhang XHD[1] Why in the Sierpiski Triangle is this set being used as the example for the OSC and not a more "natural"? the following: \[ 2 (a) The difference in sample means is an appropriate point estimate: $\bar {x}_n - \bar {x}_s = 0.40$. The only thing that differs among methods of computing the SMD is the denominator, the standardization factor (SF). In randomized controlled trials (RCTs), endpoint scores, or change scores representing the difference between endpoint and baseline, are values of interest. While calculating by hand produces a smd of 0.009(which is the same as produced by the smd and TableOne functions in R), the MatchBalance comes up with a standardized mean differences of 11.317(more than 1000 times as large. It measures the number of standard deviations a given data point is from the mean. and hit selection[2] , In the situation where the two groups are correlated, a commonly used strategy to avoid the calculation of Two types of plots can be produced: consonance 2021. WebAnswer: The expression for calculating the standard deviation of the difference between two means is given by z = [ (x1 - x2) - (1 - 2)] / sqrt ( 12 / n1 + 22 / n2) The sampling The standard error of the mean is calculated using the standard deviation and the sample size. The MM estimate of SSMD is then[1], When the two groups have normal distributions with equal variance, Leys. For this calculation, the denominator is simply the pooled standard For the SMDs calculated in this package we use the non-central To address this, Match returns a vector of weights in the weights component, one for each pair, that represents how much that pair should contribute. 2019. Can I use my Coinbase address to receive bitcoin? One the denominator is the standard deviation of P correction (calculation above). proposed SSMD to evaluate the differentiation between a positive control and a negative control in HTS assays. (which seems unexpected to me as it has already been around for quite some time). \], \[ can display both average fold change and SSMD for all test compounds in an assay and help to integrate both of them to select hits in HTS experiments Clipboard, Search History, and several other advanced features are temporarily unavailable. As this is a recently developed methodology, its properties and effectiveness have not been empirically examined, but it has a stronger theoretical basis than Austin's method and allows for a more flexible balance assessment. Standardized Mean Difference WebAbout z-scores / standard scores. Assessing for causality after genetic matching - how to use weights. deviation. multiplying d by J. If these SMDs are being reported PLoS One. Standardized Mean Difference [20][23], In a primary screen without replicates, assuming the measured value (usually on the log scale) in a well for a tested compound is The first answer is that you can't. It doesn't matter. population d. is defined as . WebThe standardized mean difference is used as a summary statistic in meta-analysis when the studies all assess the same outcome but measure it in a variety of ways (for example, all studies measure depression but they use different psychometric scales). s The SSMD for this compound is estimated as When these conditions are satisfied, the general inference tools of Chapter 4 may be applied. For hit selection, the size of effects of a compound (i.e., a small molecule or an siRNA) is represented by the magnitude of difference between the compound and a negative reference. option in the package is the nct type of confidence intervals. [20] However, in medical research, many baseline covariates are dichotomous. Zhang JH et al. {\displaystyle \sigma _{1}^{2}} Furthermore, it is common that two or more positive controls are adopted in a single experiment. \sigma_{SMD} = \sqrt{\frac{1}{n} + \frac{d_z^2}{(2 \cdot n)}} It may require cleanup to comply with Wikipedia's content policies, particularly, Application in high-throughput screening assays, Learn how and when to remove this template message, "Optimal High-Throughput Screening: Practical Experimental Design and Data Analysis for Genome-scale RNAi Research, Cambridge University Press", "A pair of new statistical parameters for quality control in RNA interference high-throughput screening assays", "A new method with flexible and balanced control of false negatives and false positives for hit selection in RNA interference high-throughput screening assays", "A simple statistical parameter for use in evaluation and validation of high throughput screening assays", "Novel analytic criteria and effective plate designs for quality control in genome-wide RNAi screens", "Integrating experimental and analytic approaches to improve data quality in genome-wide RNAi screens", "The use of strictly standardized mean difference for hit selection in primary RNA interference high-throughput screening experiments", "An effective method controlling false discoveries and false non-discoveries in genome-scale RNAi screens", "The use of SSMD-based false discovery and false non-discovery rates in genome-scale RNAi screens", "Error rates and power in genome-scale RNAi screens", "Statistical methods for analysis of high-throughput RNA interference screens", "A lentivirus-mediated genetic screen identifies dihydrofolate reductase (DHFR) as a modulator of beta-catenin/GSK3 signaling", "Experimental design and statistical methods for improved hit detection in high-throughput screening", "Genome-scale RNAi screen for host factors required for HIV replication", "Genome-wide screens for effective siRNAs through assessing the size of siRNA effects", "Illustration of SSMD, z Score, SSMD*, z* Score, and t Statistic for Hit Selection in RNAi High-Throughput Screens", "Determination of sample size in genome-scale RNAi screens", "Hit selection with false discovery rate control in genome-scale RNAi screens", "Inhibition of calcineurin-mediated endocytosis and alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors prevents amyloid beta oligomer-induced synaptic disruption", https://en.wikipedia.org/w/index.php?title=Strictly_standardized_mean_difference&oldid=1136354119, Wikipedia articles with possible conflicts of interest from July 2011, Articles with unsourced statements from December 2011, Creative Commons Attribution-ShareAlike License 3.0, This page was last edited on 29 January 2023, at 23:14. Finally, if you turn off ties by setting ties = FALSE in the call to Match, then your formula does work if you modify the standard deviation to be that of the matched treated group because all the weights in the Match object are equal to 1. Copyright 2020 Physicians Postgraduate Press, Inc. If a and another group has mean Are these two studies compatible? s (and if yes, how can it be interpreted? (Cohens d(av)), and the standard deviation of the control condition , median d(av)), and the standard deviation of the control group (Glasss $\Delta$). correct notation is provided by Lakens Makowski (2020), \[ {\displaystyle s_{N}} s [2] To some extent, the d+-probability is equivalent to the well-established probabilistic index P(X>Y) which has been studied and applied in many areas. Matching, MatchIt, twang, CBPS, and other packages all use different standards, so I wanted to unify them. material of Cousineau and Goulet-Pelletier \]. The results of the bootstrapping are stored in the results. since many times researchers are not reporting Jacob Cohens original ~ Can the game be left in an invalid state if all state-based actions are replaced? In this section we consider a difference in two population means, $\mu_1 - \mu_2$, under the condition that the data are not paired. s Matching is a "design-based" method, meaning the sample is adjusted without reference to the outcome, similar to the design of a randomized trial. {n_1 \cdot n_2 \cdot (\sigma_1^2 + \sigma_2^2)} ), Conditions for normality of $\bar {x}_1 - \bar {x}_2$. If this is the case, we made a Type 2 Error. In such cases, the mean differences from the different RCTs cannot be pooled. You will notice that match_data has more rows than lalonde, even though in matching you discarded units. To learn more, see our tips on writing great answers. ~ This QC characteristic can be evaluated using the comparison of two well types in HTS assays. The What Works Clearinghouse recommends using the small-sample corrected Hedge's $g$, which has its own funky formula (see page 15 of the WWC Procedures Handbook here). Since the point estimate is nearly normal, we can nd the upper tail using the Z score and normal probability table: \[Z = \dfrac {0.40 - 0}{0.26} = 1.54 \rightarrow \text {upper tail} = 1 - 0.938 = 0.062\]. ANOVAs., Variances Assumed Unequal: [21], As a statistical parameter, SSMD (denoted as density matrix. [24] \cdot (1+d^2 \cdot \frac{n}{2 \cdot (1-r_{12})}) -\frac{d^2}{J^2}} When applying this formula below, we see that we do indeed get the correct answer: If instead of dealing with this funky strangely-sized dataset, you want to deal with your original dataset with matching weights, where unmatched units are weighted 0 and matched units are weighted based on how many matches they are a part of, you can use the get.w function in cobalt to extract matching weights from the Match object. Standardized mean difference standardized mean differences Standardized mean differences (SMD) are a key balance diagnostic after propensity score matching (eg Zhang et al). Alternative formulas for the standardized mean difference intervals wherein the observed t-statistic ($t_{obs}$) (note: the standard error is First, the standard deviation of the difference scores are calculated. Mean Difference, Standardized Mean Difference (SMD), + Site design / logo 2023 Stack Exchange Inc; user contributions licensed under CC BY-SA. Therefore, SSMD can be used for both quality control and hit selection in HTS experiments. For example, a confidence interval may take the following form: When we compute the confidence interval for $\mu_1 - \mu_2$, the point estimate is the difference in sample means, the value $z^*$ corresponds to the confidence level, and the standard error is computed from Equation \ref{5.4}. Standardization {\displaystyle {\bar {X}}_{P},{\bar {X}}_{N}} is adjusted for the correlation between measures. rev2023.4.21.43403. -\frac{d_{rm}^2}{J^2}} the standard deviation. formulation. Please enable it to take advantage of the complete set of features! New blog post from our CEO Prashanth: Community is the future of AI, Improving the copy in the close modal and post notices - 2023 edition, Discrepancy in Calculating SMD Between CreateTableOne and Cobalt R Packages, Increased range of standardized difference after matching imputed datasets. The SMD, Cohens d(av), is then calculated as the following: \[ WebThe most appropriate standardized mean difference (SMD) from a cross-over trial divides the mean difference by the standard deviation of measurements (and not by the standard deviation of the differences). WebThe mean difference (more correctly, 'difference in means') is a standard statistic that measures the absolute difference between the mean value in two groups in a clinical \]. {\displaystyle {\bar {d}}_{i}} Language links are at the top of the page across from the title. This article presents and explains the different terms and concepts with the help of simple examples. Their computation is indeed Compute the p-value of the hypothesis test using the figure in Example 5.9, and evaluate the hypotheses using a signi cance level of $\alpha = 0.05.$. [19][22] ) of SSMD. fairly accurate coverage for the confidence intervals for any type of Strictly standardized mean difference - Wikipedia , X Shah V, Taddio A, Rieder MJ; HELPinKIDS Team. 2 WebWe found that a standardized difference of 10% (or 0.1) is equivalent to having a phi coefficient of 0.05 (indicating negligible correlation) for the correlation between treatment . For quality control, one index for the quality of an HTS assay is the magnitude of difference between a positive control and a negative reference in an assay plate. and transmitted securely. The advantage of checking standardized mean differences is that it allows for comparisons of balance across variables measured in different units. Other \]. The method is as follows: This is equivalent to performing g-computation to estimate the effect of the treatment on the covariate adjusting only for the propensity score. When considering the difference of two means, there are two common cases: the two samples are paired or they are independent. Has depleted uranium been considered for radiation shielding in crewed spacecraft beyond LEO? interface is almost the same as t_TOST but you dont set an These values are compared between experimental and control groups, yielding a mean difference between the experimental and control groups for each outcome that is compared. When the data is preprocessed using log-transformation as we normally do in HTS experiments, SSMD is the mean of log fold change divided by the standard deviation of log fold change with respect to a negative reference. attempt is significant, a researcher could compare to see how compatible Because this is a two-sided test and we want the area of both tails, we double this single tail to get the p-value: 0.124. MathJax reference. doi: 10.1371/journal.pone.0279278. K Fit a regression model of the covariate on the treatment, the propensity score, and their interaction, Generate predicted values under treatment and under control for each unit from this model, Divide by the estimated residual standard deviation (if the outcome is continuous) or a standard deviation computed from the predicted probabilities (if the outcome is binary). This means that the larger the sample, the smaller the standard error, because the sample statistic will be closer to approaching the population The SMD, Cohens d (rm), is then calculated with a None of these , the MM estimate of SSMD is, SSMD looks similar to t-statistic and Cohen's d, but they are different with one another as illustrated in.[3]. It is now clear to me and have upvoted and accepted your answer. I edited my answer to fully explain this. The standard error ($\sigma$) of Use MathJax to format equations. By default cobalt::bal.tab () produces un standardized mean differences (i.e., raw differences in proportion) for binary and categorical variables. This special relationship follows from probability theory. [20], In many cases, scientists may use both SSMD and average fold change for hit selection in HTS experiments. [10] Their computation is indeed straightforward after matching. In generic terms, the SMD n In most papers the SMD is reported as returned. The standards I use in cobalt are the following: The user has the option of setting s.d.denom to a few other values, which include "hedges" for the small-sample corrected Hedge's $g$, "all" for the standard deviation of the variable in the combine unadjusted sample, or "weighted" for the standard deviation in the combined adjusted sample, which is what you computed. Clin Ther. \], \[ n calculate the lower and upper bounds of $\lambda$, and 2) transforming this back to introduction to inverse probability of treatment weighting in Is the "std mean diff" listed in MatchBalance something different than the smd? Standard Error (type = "cd"), or both (the default option; Which one to choose? Why does contour plot not show point(s) where function has a discontinuity? {\displaystyle K\approx n_{1}+n_{2}-3.48} where $s_1$ and $n_1$ represent the sample standard deviation and sample size. The standard error ($\sigma$) of \] wherein $J$ represents the Absolutely not. d 5. Differences between means: type I As it is standardized, comparison across variables on different scales is possible. Why is it shorter than a normal address? or you may only have the summary statistics from another study. n_2(\sigma^2_1+\sigma^2_2)}{2 \cdot (n_2 \cdot \sigma^2_1+n_1 \cdot are the medians and median absolute deviations in the positive and negative controls, respectively. \[ To learn more, see our tips on writing great answers. Thanks for contributing an answer to Cross Validated! . Find it still a bit odd that MatchBalance chooses to report these values on a scale 100 times as large. and median absolute deviation Goulet-Pelletier 2021). {\displaystyle n} Pediatrics. Ferreira IM, Brooks D, White J, Goldstein R. Cochrane Database Syst Rev. As a rule of thumb, a standardized difference of <10% may be considered a The degrees of freedom for Cohens d is the following: \[ Indeed, this is an epistemic weakness of these methods; you can't assess the degree to which confounding due to the measured covariates has been reduced when using regression. Healthcare Utilization Among Children Receiving Permanent Supportive Housing. The degrees of freedom for Cohens d(rm) is the following: \[ s_{diff} = \sqrt{sd_1^2 + sd_2^2 - 2 \cdot r_{12} \cdot sd_1 \cdot In theory, you could use these weights to compute weighted balance statistics like you would if you were using propensity score weights. cobalt provides several options for computing the SMD; it is not a trivial problem. specify goulet (for the Cousineau and Thanks for contributing an answer to Cross Validated! Standardized Difference Because each sample has at least 30 observations ($n_w = 55$ and $n_m = 45$), this substitution using the sample standard deviation tends to be very good. and sample variance section. ~ The limits of the t-distribution at the given alpha-level and degrees Use MathJax to format equations. Federal government websites often end in .gov or .mil. We found Using this information, the general confidence interval formula may be applied in an attempt to capture the true difference in means, in this case using a 95% confidence level: \[ \text {point estimate} \pm z^*SE \rightarrow 14.48 \pm 1.96 \times 2.77 = (9.05, 19.91)\]. The different ways of computing the SF will not affect its value in most cases. at least this large, ~1% of the time. The 99% confidence interval: \[14.48 \pm 2.58 \times 2.77 \rightarrow (7.33, 21.63).\]. There may be a few other weirdnesses here and there that are described in the documentation. { "5.01:_One-Sample_Means_with_the_t_Distribution" : "property get [Map MindTouch.Deki.Logic.ExtensionProcessorQueryProvider+<>c__DisplayClass228_0.b__1]()", "5.02:_Paired_Data" : "property get [Map MindTouch.Deki.Logic.ExtensionProcessorQueryProvider+<>c__DisplayClass228_0.b__1]()", "5.03:_Difference_of_Two_Means" : "property get [Map MindTouch.Deki.Logic.ExtensionProcessorQueryProvider+<>c__DisplayClass228_0.b__1]()", "5.04:_Power_Calculations_for_a_Difference_of_Means_(Special_Topic)" : "property get [Map MindTouch.Deki.Logic.ExtensionProcessorQueryProvider+<>c__DisplayClass228_0.b__1]()", "5.05:_Comparing_many_Means_with_ANOVA_(Special_Topic)" : "property get [Map MindTouch.Deki.Logic.ExtensionProcessorQueryProvider+<>c__DisplayClass228_0.b__1]()", "5.06:_Exercises" : "property get [Map MindTouch.Deki.Logic.ExtensionProcessorQueryProvider+<>c__DisplayClass228_0.b__1]()" }, { "00:_Front_Matter" : "property get [Map MindTouch.Deki.Logic.ExtensionProcessorQueryProvider+<>c__DisplayClass228_0.b__1]()", "01:_Introduction_to_Data" : "property get [Map MindTouch.Deki.Logic.ExtensionProcessorQueryProvider+<>c__DisplayClass228_0.b__1]()", "02:_Probability" : "property get [Map MindTouch.Deki.Logic.ExtensionProcessorQueryProvider+<>c__DisplayClass228_0.b__1]()", "03:_Distributions_of_Random_Variables" : "property get [Map MindTouch.Deki.Logic.ExtensionProcessorQueryProvider+<>c__DisplayClass228_0.b__1]()", "04:_Foundations_for_Inference" : "property get [Map MindTouch.Deki.Logic.ExtensionProcessorQueryProvider+<>c__DisplayClass228_0.b__1]()", "05:_Inference_for_Numerical_Data" : "property get [Map MindTouch.Deki.Logic.ExtensionProcessorQueryProvider+<>c__DisplayClass228_0.b__1]()", "06:_Inference_for_Categorical_Data" : "property get [Map MindTouch.Deki.Logic.ExtensionProcessorQueryProvider+<>c__DisplayClass228_0.b__1]()", "07:_Introduction_to_Linear_Regression" : "property get [Map MindTouch.Deki.Logic.ExtensionProcessorQueryProvider+<>c__DisplayClass228_0.b__1]()", "08:_Multiple_and_Logistic_Regression" : "property get [Map MindTouch.Deki.Logic.ExtensionProcessorQueryProvider+<>c__DisplayClass228_0.b__1]()", "zz:_Back_Matter" : "property get [Map MindTouch.Deki.Logic.ExtensionProcessorQueryProvider+<>c__DisplayClass228_0.b__1]()" }, [ "article:topic", "authorname:openintro", "showtoc:no", "license:ccbysa", "licenseversion:30", "source@https://www.openintro.org/book/os" ], https://stats.libretexts.org/@app/auth/3/login?returnto=https%3A%2F%2Fstats.libretexts.org%2FBookshelves%2FIntroductory_Statistics%2FBook%253A_OpenIntro_Statistics_(Diez_et_al).%2F05%253A_Inference_for_Numerical_Data%2F5.03%253A_Difference_of_Two_Means, $ \newcommand{\vecs}[1]{\overset { \scriptstyle \rightharpoonup} {\mathbf{#1}}}$ $ \newcommand{\vecd}[1]{\overset{-\!-\!\rightharpoonup}{\vphantom{a}\smash{#1}}} $$\newcommand{\id}{\mathrm{id}}$ $ \newcommand{\Span}{\mathrm{span}}$ $ \newcommand{\kernel}{\mathrm{null}\,}$ $ \newcommand{\range}{\mathrm{range}\,}$ $ \newcommand{\RealPart}{\mathrm{Re}}$ $ \newcommand{\ImaginaryPart}{\mathrm{Im}}$ $ \newcommand{\Argument}{\mathrm{Arg}}$ $ \newcommand{\norm}[1]{\| #1 \|}$ $ \newcommand{\inner}[2]{\langle #1, #2 \rangle}$ $ \newcommand{\Span}{\mathrm{span}}$ $\newcommand{\id}{\mathrm{id}}$ $ \newcommand{\Span}{\mathrm{span}}$ $ \newcommand{\kernel}{\mathrm{null}\,}$ $ \newcommand{\range}{\mathrm{range}\,}$ $ \newcommand{\RealPart}{\mathrm{Re}}$ $ \newcommand{\ImaginaryPart}{\mathrm{Im}}$ $ \newcommand{\Argument}{\mathrm{Arg}}$ $ \newcommand{\norm}[1]{\| #1 \|}$ $ \newcommand{\inner}[2]{\langle #1, #2 \rangle}$ $ \newcommand{\Span}{\mathrm{span}}$$\newcommand{\AA}{\unicode[.8,0]{x212B}}$, 5.4: Power Calculations for a Difference of Means (Special Topic), David Diez, Christopher Barr, & Mine etinkaya-Rundel, Point Estimates and Standard Errors for Differences of Means, Hypothesis tests Based on a Difference in Means, Summary for inference of the difference of two means.