The purpose of this study was to compare how students in Grades 3 to 6 respond to a mathematics attitudes instrument with a 4-point Likert-type scale compared with one with an additional neutral point (a 5-point Likert-type scale). The 606 participating students from six elementary and middle schools randomly received either the 4-point or 5-point format of the Math and Me Survey. Regardless of whether a neutral midpoint was offered or not, the structure of the instrument was virtually the same, with equal intercepts, means, variances and covariances, pattern coefficients, and nearly all residuals. The 5-point scale is preferred with this population because with this format the reliability estimate for the Mathematical Self-Perceptions subscale was higher (p = 0.049), and the pattern coefficients were stronger. Additionally, this format provided less model misfit than the 4-point format. Based on these findings, the authors recommend administration of the Math and Me Survey in the 5-point format. These findings also indicate that despite what some educators and educational experts believe, children in Grades 3 to 6 are capable of discriminating among five response options and do not tend toward the neutral point more so than with a 4-point scale. (Contains 2 tables.)