企業成長決策取決於企業自身之資源及其管理人對外部機會的洞察力，因此決策本身並非隨機而是由管理人選擇其偏好的策略，即成長決策本身存在潛在的自我選擇偏誤，若實證模型忽略廠商自我選擇偏誤的影響，將使實證結果產生偏差。因此，本研究採Heckman二階段決策模型，以控制自我選擇偏誤的影響。第一階段，係以690個上市（櫃）電子資訊業樣本建立Probit模型，以檢驗企業之剩餘資源是否影響其成長決策；在第二階段納入自我選擇修正因子，以OLS迴歸分別建立持續成長與中斷成長二組樣本之成長－績效模型。實證結果發現財務資源對企業持續成長之機率呈倒U型曲線關係；無形資源與企業持續成長之機率則呈U型曲線關係。第二階段，在控制自我選擇修正因子後，企業成長與績效呈顯著正向關係。

A firm's growth decision depended on its own resources and manager given its insight into the opportunities from external environment. Therefore, the firm's growth decision is not made at a random, but is mangers to self-select into their preferred strategies. In other words, there is a potential of self-selection bias in growth decision. If the self-selection bias is not taken into account then any conclusions drawn may be invalid. Hence, this paper controls self-selection bias via Heckman's two-stage decision model. The first stage, this paper utilizes a Probit model to examine a firm's slack resources whether or not influenced its growth decision for 690 samples of publicly traded electronics industry. And the inverse Mills ratio that corrects for self-selection matters in the second stage OLS regression for both continuous growth and setback growth samples. The empirical results of the first stage, we found that the relationship between financial resources and the probability of continuous growth is a U-shaped relationship. In additon, the relationship between intangible resources and the probability of continuous growth is an inverse U-shaped relationship. The second stage, we found that there is a positive relationship between firm growth and performance when the self-selection bias has been controlled.